Hi Leo, back in the day, when CLI was a the only thing out there, PGP for email had caught on. I used to use a product called Eudora and similar email implementations. It seems like it took a while for this to come back in a different form. Better late than never.
@@debtfordwharf It never really went away. There was always a plugin for Thunderbird, and assorted others. The problem is it's still too complex for normal people to use.
If someone adopts Passkeys, should they delete all other methods of authentication they used previously? For instance, Google Prompts? Could someone exploit/intercept Google Prompts if used at some point despite the fact that we set up Passkey?
Passkeys are more secure than passwords because they are less "powerful". With an username/password pair you can potentially log in from any account, on any device, anywhere. Anyone who gets your username/password can potentially masquerade as you from anywhere. That is what makes them less secure / more dangerous. As described in the video, a given passkey is tied to a specific account and a specific device. It is important to remember that passkeys only authenticate the device and account. They do not authenticate the person. This is why the total security solution requires you keep access to the device secure. Fingerprint scanners, FaceID, or PINs HAVE to be used so that people who have physical access to your device, can't actually access your account. The device/operating system you use must, of course, provide a mechanism for securely operating a keystore. This was an excellent video.
This is a naive point of view. Passkeys will be stored in raw form 100% it's impossible to guarantee biometrics on everything and syncable is a core requirement for an everyday user. Because of that you open the opportunity to steal the passkeys from a device guaranteeing that a successful attacker will have access to every. single. account. ever. forever. As you have correctly stated, you have moved away from authenticating people to authenticating just the device. This is an inherent risk introduction.
this is the clearest, most understandable explanation of what a passkey is on youtube, or anywhere else. and it also explains that apps are actually using passkeys when they allow us to login using our fingerprint, face, or pin. and also an added bonus on what are pgp keys. im so glad i watched this. thank you very much.
9:01 this is where Steve Gibson’s SQRL protocol is superior to passkeys. Both use public key encryption, but SQRL has 1 identity that creates a key pair on the fly for each login based on the site’s domain name. Elliptic curve crypto allows you to create a private key based on a determined input. The same input will always create the same key. Therefore a secret (the identity) mixed with the domain name will create a unique key pair for each login. Since this is easily calculated, there’s no need to save it for each site, just keep the original secret (identity)and recalculate based on the domain. This means the protocol and any devices can have an unlimited number of sites to log into, no extra storage and it’s easily shared between devices. Oh well, we get passkeys instead.
I guess there is no golden bullet in cryptography. Each solution has it's strengths and weaknesses, but what is considered a strength (or weakness) by one user might be the opposite for another user. Even though the passkey in it's current form is perhaps less elegant and definitely more cumbersome than the method you describe, I would still prefer the current solution. If I understand you correctly, I personally wouldn't like to have solution like SQRL that is based on a single private secret used to authenticate/unlock all (!) my accounts/logins. Sounds like the likelihood somebody would be able to crack my private secret in the future would increase with the number of accounts/logins and of course advances in technology with time.
Wow yeah, I second that. You have a very easy to understand way of explaining this!! Thanks so much. Every time I think I understand the encryption/decryption process, I seem to lose the understanding. This helped immensely.
This is an amazing explanation. Thank you for making it so clear. I will be saving this video so that when anyone asks about passkeys, I will share this to them.
I consider myself a pretty savvy techie. I’ve always understood, encryption and public key and private key stuff. But for passkeys, I have some sort of mental block. That being said, Leo‘s explanation of passkeys is by far the best I have ever heard!
I think because all other explanations contain the "and then magic happens" part, this is the first one that explains that passkeys are just ssh-like authentication with a better UI.
What i like about things like this is that they are complicated for most users and this causes things to go wrong, so you end up dropping down to passwords and email to get back in to most accounts.This negates the purpose of it. you basically bypass by clicking on the "I forgot my password" link, this mostly ends up going back to unsecured emails.
As ever Mr LN explains the inexplicable with ease. Been a follower for years - when internet connections required a series of morse-code-like noises and then went at speeds the common tortoise scoffed at
In a passkey-only service, isn't there a higher than normal risk of getting blocked of your own account if you lose the devices the passkey is stored on?
I see the convenience of passkeys, but for me there's still a problem. Passkeys are effectively single factor authentication. Mere possession of the passkey is generally enough to gain access to a passkey protected system. If a criminal steals your laptop and gains access to it (e.g. by shoulder-surfing your laptop password), then they can automatically access any passkey protected data you have. Using complex passwords protected by a password manager (with a strong master password) together with 2FA (using a password protected OTP generating app), whilst much less convenient, seems far more secure. In the event of a data breach at say your email provider, even if hackers got access to your email password, 2FA would still prevent them from accessing your mail.
+1 also agree with my current understanding, but I guess passkeys is still an improvement for the "below average" user that we realistically have given up on trying to make them all put in the effort to adopt the password manager + 2FA setup
No. No. No. In the description you give here, your laptop password is your first factor and the physical passkey is your second factor. That's 2FA. If you let both get stolen, it was still 2FA. If you leave your passkey in your laptop fulltime, then it's no better than using an authenticator app installed on your laptop. A passkey only becomes stronger than a laptop authenticator app if you only plug in your passkey during login and then you physically remove it and store it where it cannot be stolen simultaneously with your laptop.
A passkey is not a hardware key that you can plug and unplug from your computer. You might be thinking of a yubikey? A passkey is a cryptographic hash that stays on your device
You need a devices with fingerprint or facial authentication to create a passkey. So if your computer doesn’t have biometric authentication, you will need your phone to create and store the passkey. You’ll then confirm the access from the computer using your phone. The system is well thought and has been developed by Apple, Microsoft and Google. Nowadays all modern OS’s are compatible. The biggest problem is passkey synchronization between devices so that you don’t loose all your accesses if you loose your device. Apple has the keychain for that, Microsoft and Google will surely also have an equivalent.
Thank you! This is the best explanation of passkeys I've heard so far! One of my concerns regarding passkeys is... what happens when you have an account that's only using passkeys, have only setup passkeys for that account on a single device, and that device is lost, stolen, or is otherwise unavailable (it dies)? How do you regain access to that account? It seems like the best defense for such a situation is to have passkeys setup on multiple devices, allowing you to confirm you identity when setting up a new device after a device becomes unavailable, but that's not economically viable for some people. An alternative is to actually have a password for the service, using passkeys when possible, but that leaves the account vulnerable in the event of a data breach. Additionally, let's say I want to replace a functional device (my only device) with a new device. It seems I would need to maintain possession of that device for some "overlap period", during which I would need to login to every service I use on the new device, so that my new device can be authenticated by the old device. That seems rather cumbersome, but is probably a small price to pay for the added security of passkeys. What are your thoughts?
Each time you set up a passkey on a new machine a different form of authentication is used. For example a code to your phone, or a message to your email. Once set up it becomes your authentication mechanism. But you're always able to set it up from scratch somehow.
This is an excellent question and one that bothered me for a while. You can’t make the argument that you are in a better security position with passkeys if the use of passkeys is in addition to an authentication method that was already present. Therefore, you have only improved your security posture if you remove the old auth method and only use passkeys. However, if you do this, you run into the issue you are asking about. I think for this scenario is exactly why having a 3rd party password manager (PM) in general, and 1Password in particular, makes sense. The PM collects and manages all the passkey private keys so no matter what happens to the device that actually created them, it doesn’t matter. You get your new phone, authenticate to the PM, and you are back in business. But now isn’t the PM vulnerable? Not with 1Password’s security architecture. There are two necessary pieces of information to access the 1P vault that are never stored in the cloud or even transmitted: your password AND a locally generated random security key. You pair those things with a hardware security key, stored in multiple secure locations, and I think you have a setup that’s nearly impossible to breach, but is also convenient
Well, I think passkeys are just a convenience mechanism in that you have to authenticate only once either in the key manager of your OS or in your password manager and then use the per device generated and stored passkeys to log in to the websites. No need to manage different passwords, and it also increases security as you are not exposing your password in your daily login routine. No chance of some man in the middle or some other malicious browser extension stealing your password. Now your concern about a data breach happening on the website on which you use the password to login, most of the companies don't store raw passwords in their databases. They store salted one way encrypted password. As soon as you supply the password and try to log in, it is immediately encrypted in the client side and transported to the server in an SSL tunnel ("s" in HTTPS indicates that the site uses SSL, which means all traffic is encrypted)
This is a much better and clearer explanation than the one provided by Google. However: - at my last work place (a major corporation) using ssh keys was not allowed by Info Security group. According to them it violated their security standards. - if using passkeys prevents me from accessing my Google account from a device which doesn't have my passkey, then I have to use password, which defeats the purpose - if my passkey is automatically created on a new device, what if I'm using a public device and don't want my passkeys on it? - with passkeys I'm relying on a strong login security of my device, but if someone breaks in into my device, then they have green light to all my accounts
1) your IT folks are misguided. :-) 2) We're moving to a passwordless future. In some cases you can choose that right now, in others you can simulate. So, no it does not defeat the purpose. 3) Passkeys will never be created without asking you first. 4) No. They'd still have to pass biometric or Windows Hello authentication when they attempt to use a passkey.
Leo, The major concern I have regarding the password-to-passkey transition period, is that the service/company/app I am accessing will actually have both the new public key for a specific device(s) AND my original password. In the example you used where the service was hacked and they stole my public key, e-mail, etc., didn't they also get my still usable password? I mention this because I have created a few passkeys but have not seen an option to have the service permanently delete my password once the passkey was created. Therefore, even if I create or share passkeys for all my devices to a particular service, a data breach of that service will cause the same pain it does with or without passkeys because my passwords are stored in the same old way "alongside" my public key. What am I missing? Thanks for the excellent video! Karl
Microsoft allows you to remove passwords. In leiu of that, set the password to something ridiculously long and complex, and then don't save it anywhere. Nothing for hackers to steal, you'd never use it yourself, so it's as close to password-free as you can get.
@@askleonotenboom Thanks Leo. (I am in the Apple ecosystem.) Unless I misunderstood, which is entirely likely, the service still has my valid user ID and/or email and my long/complex password. Many companies use very poor data security practices and leave files exposed on cloud servers like it is a hobby, so I still believe the danger is being unable to have passwords removed from a service when you have switched entirely to passkeys. BTW, although companies frequently use poor data security practices, they are expert at apologizing after they have been breached. I currently have three different "free" identity theft services due to breaches. Being breached does not seem to cause the same reputation damage it used to. Keep up the good work! Karl
Have been following passkeys for a while but have never seen such a clear explanation. Congratz! Regarding passkeys i do have 2 concerns 1. Suppose i loose my device with the only private key i have, how will i be able to restore my account on a new device? 2. When creating a passkeys for an existing account, the less safe login method using a password which could be stolen from the server still exists. Hope some one can convince me that both issues can solved.
When you set up a passkey on a new device, yes, you login some other way. It could be password, but it need not be. It's more often something more secure like a confirmation email sent to the email address of record, or a text message to the phone number of record, or similar. Once you've confirmed your identity that way, the passkey is created. Losing your device has nothing to do with any of that. ANY new device on which you want to set up a passkey goes through that process. If you lose your device, however, once you've signed in to the account elsewhere you can remotely disable the passkey associated with that account.
The public key would be like walking down a street and writing down the house numbers you see on mailboxes, but that won't unlock the deadbolt on the front door...
So, no more passwords or password managers? What about the security of the pin? Do you need to use different pins on different devices? If I were to lose my device, wouldn’t the pin be easily compromised? Or would pins therefore need to be treated as passwords are now? My passwords are 20+ characters long & in a password manager. I question that use of passkeys & a pin is better than what I currently do. Plus, is it easy to change the passkey if a device is changed? I assume removing a device is done by wiping the data from it. But, the old passkey for the old device needs to no longer be valid.
"no more passwords or password managers?" -- Not necessarily. People may choose to use their favorite password manager to manage their private keystore. "What about the security of the pin?" -- The PIN is handled locally. If you have the device and you know the PIN, then you're good. Unless someone has your device, knowing the PIN alone is useless. "Do you need to use different pins on different devices?" -- Each device is independently secured, so that would be the user's choice. "If I were to lose my device, wouldn’t the pin be easily compromised?" -- If you lost your device, then anyone in possession of the device could access your keystore if they know your PIN. Obviously, a PIN would be easier to crack than a 20+ character password, so best not lose your device. "I question that use of passkeys & a pin is better than what I currently do." -- Passkeys provide some significant advantages, but they have disadvantages as well. You need to choose an appropriate mix of risks that work for you. Passkeys give certain protections against phishing attacks, but are subject to being lost when the device is lost, inaccessible, or nonfunctional. Incorporating passkeys into a password manager service provides some safety here, but then there's the question of how you intend to log into your password manager. " is it easy to change the passkey if a device is changed?" -- That will depend on the site being logged into, and whether or not you plan on managing your passkeys with a password manager service. Using a password manager provides excellent portability between devices: changing a passkey on one device would affect all other devices as well. I am confident that changing a passkey on a given site won't be a terribly difficult issue, but who knows? We still have places where your password length is limited to 16 characters (or less). "But, the old passkey for the old device needs to no longer be valid." If you want to invalidate a passkey (if it's tied to a device, instead of a password manager), that would be done on a site-by-site basis.
@@neuideas John Cole One query If a passkey can be uploaded on a online password manager and can be used to login on a new device, then How come is it different from a password mechanism ?? Also, accessing password manager using Local pin, fingerprint or face lock is not possible since its a new device ?? I still cannot understand how I will login on a new device if its lost if I own only a single device
@@himanshuchhabra1942 " is it different from a password mechanism ??" It's different because it uses public key verification of digital signing, rather than hashing a password. The burden shifts from remembering your login information, to not losing your private keys. The private keys are never revealed, so they are never risked. "Also, accessing password manager using Local pin, fingerprint or face lock is not possible since its a new device ??" You will need to authenticate the application first. This can be done by validating your account on an already authenticated device. "I still cannot understand how I will login on a new device if its lost if I own only a single device" You need to set up a backup device first, before you lose your primary device. Alternately, you may be able to log in using one-time use backup codes, if they are offered. Beyond that, you could always fall back on standard login credentials, assuming that's an option.
@@neuideas I understood the mechanism , I was trying to understand the big picture. My point was private keys revelation is not needed, if the hacker can use other ways of authentication which are provided for the user in case he loses the device.
Leo, this was an absolutely superb presentation. I already had a decent understanding of private/public key pair but I knew nothing about passkey. I'm going to see how my very non-technical wife can follow your video now :-)
Great explanation. From a privacy perspective, I’ve always wondered about adding a fingerprint or face as a ‘passkey’. Is this stored anywhere that compromises one’s privacy?
Thanks for your nice overview Leo. I’m interested in how third party password manager apps will help manage this information, versus the device operating system itself.
3rd party password apps aren't necessary. The key pairs are known only to your computer (the private key) and the site you're accessing (the public key). No password for 3rd party password apps to manage.
@@HarshColby Some online 3rd party password managers can store your private key. That's how they sync your passkey between all devices where the password manager works. Some of those online apps include 1Password and Bitwarden. KeepassXC is an offline password manager that will soon support passkeys, too.
Thank you. You described a software pair of keys. And Google does supply that. But, there are vendors selling hardware devices. I assume that using hardware PassKeys, the public key is identical on each of the web services that I use the hardware PassKey. When should I consider buying the hardware PassKey? Do most web services also require a password in addition to the PassKey?
Actually if the yubikey can provide user verification via a pin, fingerprint, or something like that, it’s considered a passkey! Most people use them for 2SV, but they’re very much usable for passkeys (given they can perform user verification)
@@askleonotenboom Yubikey 5 series can be used for passkey (FIDO2/Webauthn) authentication. Actually, that seems to be the only (simple) way currently to use passkey authentication on Linux (not Android) devices.
Impressive video. If anything, the private key is the weak link. So I am left with the doubt that the private key is safe. I know that if one has possession of the hardware with linux the logon procedure is not going to be much of protection. The only protection I would trust in that case is both disk encryption and a logon password. And make sure to switch your computer off, or someone might add a password to the list of passwords for disk encryption (LUKS). However, even that has a shelf life as quantum computing is around the corner. Of course the whole encryption scene will change by that time. Anyway, I thank you for your explanation sir, very clear!
If I create the passkey on my device (smartphone) and a private key is generated based on data from my device and my biometric data, the only way to compromise a passkey-protected account would be to hack my password manager ? Or did I not understand well? The big difference is that I, the user, do not know my private key as I do for the password as it is a very long and complex alphanumeric string.
Great video Leo. I can't wait for passkeys to take over the password phase. Do you have a list of services that have already started using passkeys, besides google?
Gee, I finally know about passkeys. I was so curious about them. One problem, however: if someone breaks into your house, and if you are not there and your computer is turned on, they can just sit down at your computer and login anywhere, can they not? Maybe the operating system would ask them for a pin, or a fingerprint.....
@@viktorpaulsen627 Not really, no. It's closer to a plain old password replacement that's more secure. Kinda. Some think of it as 2FA because your device will prompt you for your PIN/fingerprint/face before providing a passkey, but that's still only one factor that you had to provide in the moment.
Hmmm, thanks for the great explanation . I wish google would hire you to explain. Dreadful articles. However I am already deep in the swamp of questions. Such as- your laptop has problems any you have to give it to a repair person (not a hard drive replacement). They need admin rights. Does that leave you open? Also, another question - in an emergency (you can’t give them data) how do they get into your device to pull data? Lots and lots of questions before the complicated users feel comfortable. Again, thanks
"Does it leave you open" yes. Choose trustworthy repair people. (Or take extra steps to secure sensitive data while still leaving the machine operable.) Not understanding the emergency scenario you're describing, though.
@@askleonotenboom But if it's possible to sign in some other way doesn't that lower the level of security? I apologize for seeming dense but considering the skills of those with less than honorable intentions and the amount of information stored in the cloud this system may be an improvement on the current model but it still isn't perfect. I suppose anyone with these concerns could just have two machines with access then if one died you wouldn't be stuck. It will be interesting to see how it will all work out in the end. Thanks for your prompt reply.
So, it should be strongly recommended to generate passkeys from more than one device for each account, inmediately after creation or activation of passkeys.
@@mfr2 Not necessarily. It depends on the service, but like I said, you probably signed up with an email address so an email to that address could also confirm you're you.
@@frederickclause2694 Of course it's not perfect. There's no such thing. But it is significantly more secure than password based authentication. AND it's easier to use. 🙂
As a senior citizen I am very concerned that when something happens to me, my children can access all my accounts and information. If I set up a passkey, do they have to have my device to do so?? Right now, I keep passwords in an encrypted file for which they have the password and I send them the current file on a semi-regular basis. This sounds wonderful if YOU are the only one using your device and the only one needing to log into accounts. But I am a bit confused about how it will work in a situation where multiple people need to access the same account (a bank is an example).
Where does PGP email encryption, decryption, digitally signing emails fall in the mix? Sorry, I am not a techy, just someone trying to learn and understand. Thanks.
Hello.Thank you for this very interesting and informative video. A question of security: I imagine that the public key and the private key are created using an algorithm that ensures the link between the 2. What happens if a hacker gains access to this algorithm? Can he decrypt the private key? This is a very unlikely hypothesis and the risk of ordinary passwords is certainly much greater.
Now that TSA and flight security systems around the globe and immigrations check points are using fingerprint and facial identity, what could possibly go wrong? I would think that a pin number would be a better choice for the final authentication, while using these passkeys.
What if I have only one device storing my private key (say my phone) and I lost it. What is the recovery option here? and the person who was able to get my lost phone figured out the login pin? I am not questioning the security of passkey but trying to understand this scenario. In case of password, I know it and can use it from any device. By the way excellent explanation!
There is ALWAYS another way to sign in. That other way may involve more steps and be less convenient (say, emailing you a code), but think about how you establish a passkey to begin with: you have to login somehow. Once you're logged in you can then revoke the passkey assigned to your phone.
If the site owner isnt forced to update to passkeys it will take ages for this to be implemented. And if its mor conveinient can be discussed. I have setup passkey on my google account. So everytime i log in i have input my win passcode. Compaired to chrome just remembering the password. Not easier. Or i have to buy a camera or a fingerprint reader for my desktop. And how do i know if my fingerprint/face dont get stolen by hackers exploiting flaws in the camera/fingerprint reader? Imo just a new set problems compaired to the old way.
Does the rollout of pass keys present a vulnerability for accounts that have not been set up yet that may be compromised, allowing a threat actor to set up the pass key on another device? Does the key have a method of being reestablished manually or during a password reset or account recovery process?
thnx. one question, what i missed explaind. say you have done so for your phone. all is working fine. you now get a new phone. do you have to start all over again on your new phone for every account (as where it a other device as you mentioned) , and then can without a problem get access to that account ? just wondering....thnx for the feedback
Depends on the system you're signing into. Worst case you start over, but in general it could be as simple as a one-time additional hoop to jump through (a text message to confirm, and email to respond to, or another device on which to approve the sign in).
It would be nice if the user could decide which data has to be entered each time they log in. I would use username, password, 2FA and passkey together. The passkey would protect against an attacker being able to log in with my remaining login data from someone else's computer.
The point of potential confusion is where the private key is going to be stored. It’s not obvious. If you live in a multi platform environment where you login to website from various OS and mobile devices running different OSes, who is the private key store cross platform? Do I use my Google account? Do I use my 3rd party password manager? My 2 or 3 OSes? I’m already concerned that Google Authenticator offers to backup your 2FA keys in the cloud, thats putting all your access keys in one basket and outside of your physical control. I only use the QR export and import to a separate device not via the cloud. If I put passkeys in 1Password, what if my account is hacked? All the eggs are in one basket. PKI relies on the private key not getting duplicated outside of your control. This is the weakest link if we allow the private keys out into the cloud.
Each platform gets its own, unique, private key the first time you sign in, and it's stored in that platforms secure store. Tools like 1Password can act as that store to make passkeys available across platforms if you like.
@@askleonotenboom "In theory". As I've been testing passkeys cross platform (including Yubikey, Google and Microsoft Authenticators), I've yet to get the 1password passkey for my Adobe account (as a a test) to work x-platform. My android 14 phone and iPad OS 17 aren't accessing 1password for the passkey. The xp functionality still needs a lot of work.
My question is, what happens if you create a passkey and use it on your mobile device for your crypto account. Is there any instance where you could lose any kind of access to your passkey, to where you couldn't access your crypto account or mobile device?
I saw this on my PlayStation account I have lost accounts before so I was looking for a different way to keep my account safe but I didn't understand it thank you for this video
Okay, I know and understand that passkeys > passwords, but I still have a few questions. Your house burns down and with it your laptop, desktop, tablet, and phone. How do you recover those passkeys? Now I understand how this can be done, but so will the black hat hackers in Elbownia. What can we do to prevent that attack? Next, what about brute force on the public key? Yeah, it would take a zillion years to do it on a laptop today. We had the same thing with the Windows NT password system back in the 90s, and just a few (like 4-5) years laters, it was possible to crack within hours using a desktop computer. With all the computing power on earth today (equal to a single GPU in 2050), how long does it take to get through all the possible combinations? Are there collisions (you create an A-B pair, but it turns out that R can also Decrypt some A? Can there be a 'dictionary' of passkeys? There are a number of formulas and prime numbers in use. Once they are fed through, is it possible to recognize the A key and then pull the B key out of a dictionary as well?
Regarding the house burning down question. If you're in the Apple ecosystem, your passkeys are stored on your iCloud Keychain. If all of your Apple devices are destroyed in the fire, when you go to set up a replacement device you'll need to go through their Apple account recovery process (assuming you lost the appleid/password in the fire too). It's a good idea to prepare for this beforehand (for example, you can nominate someone else's account as your recovery buddy). I believe Google has a similar password manager these days.
Great explanation and examples, but I have a question! Let say that we are in 2027 all my accounts are using passkeys and there are no passwords anymore and all the private keys are stored in my machine. Now, let's imagine that someone stole my machine or it gets damaged, how can I access my accounts again without my private keys?
Your passkey is never the only way authenticate. Remember, in order to set up the passkey on each device you had to authenticate some way that didn't involve a passkey. So that continues to work when you move to a new device. A common password-free example is a link sent to your email account.
They must have "both public and private key codes" in the file or operating system. Otherwise system cannot proceed with the matching process. Hackers can find the way steal the private key code. By far the best method is to send a temporary code to your private device. Use it as your private code.
This is a superb explanation. I have one dumb question: is there no way for the server side and the device side to be hacked at the same time, or even at different times, and somehow allow the private and public keys to be mated up by the hackers?
Some password managers like 1Password will store your passkey for use elsewhere. Otherwise, passkeys are per-device, so you'd set one up on each device the first time you sign in to the account on each device.
The same way you got into the account when you first setup a passkey on that machine. Usually a different, more cumbersome, authentication process, like an email to your email address with a code or something similar.
Let's say passkey is the only way to login to an account. If you have it on one device only, then losing the device = loosing account. Also if you just happen to go on a trip and forget that device with the passkey then you can't login. I sort of prefer to not be blocked from my account like that
Good clear explanation. My concern is that each device is locked to each passkey. This potentially could be a lot to manage. Also a source of privacy issues. Given the exponential growth of surveillance capitalism I will avoid use of passkey until it's better addressed.
Not much to manage. Registering a device requires a click and is thereafter transparent. Apple and Google😢 offer the option to store the private keys in the cloud so that one key-pair can be used across multiple devices. The choice depends in part upon which you trust more, your device or the cloud.
My understanding is that if I want to use another device to check my email, it had to be authenticated. Each key pair uses the device ID as part of the encryption. Not normally a problem, but if I want to use my friends or public computer that might be an issue.
Passkey, it is the server provide a challenge to the user to be signed by private key or server signs a challenge and ask the client what the challenge is?
In the context of Passkeys, the server typically provides a challenge to the user (or client) that must be signed using the user's private key. Here’s a brief overview of the process: Challenge Generation: The server generates a unique, random challenge (a nonce) and sends it to the client (user). Signing the Challenge: The client uses their private key to sign the challenge. This ensures that the client possesses the private key without revealing it. Response to the Server: The client sends the signed challenge back to the server. Verification: The server verifies the signed challenge using the public key associated with the user's account. If the signature is valid, it confirms that the user has successfully authenticated.
My main concern about this is that I can't easily imagine what could be a smooth transition to a new device (computer or phone for example) when the previous one is damaged or simply obsolete.
@@askleonotenboomThe first time it's one site at a time and progressively. When switching computer of phone with hundreds of accounts already registered with the previous device (I've more than 800 entries in the "personal" section of my KeePass database) it could be not so trivial.
I'm trying to figure out what happens if you lose access to the device which authenticates the credentials.... like the fingerprint sensor on my phone ... if my phone explodes or gets stolen... how am I going to gain access to my account?
It's a good explanation, but I am going to create a fictitious email account and set up a passkey with a fingerprint. Then I will try to access the fake account via a desktop that doesn't have a fingerprint scanner. This is where I have trouble seeing how the same account can be accessed by a phone, and desktop and a Chromebook without f*rting around with all sorts of different pins.
I may be a novice on this topic, but I thought that authentication with a private/public keypair was a matter of signing a "nonce" (a randomly generated number) with your private key, and the server using the public key to verify the signature. Am I wrong?
I believe you're correct. What you're referring to as a nonce is essentially the "I'm thinking of a number" referenced in the video. I did say it's a very high level overview and abstraction. 😀
Of course. You had to setup the passkey on that device by authenticating some other way to begin with, so you would simply authenticate some other way on a new device and setup a new passkey.
Let me see if I understand. With password, both clients and server store the pre agreed password. The exchange between the two is done by creating a secure channel with public and private keys, and then I send the password from client to server to get authenticated. With passkey, we still create a secure channel using public and private keys, but this time the server sends a challenge (say random number) to client. Client authenticates (say face recognition) authentication done on client side. Then it sends the challenge back to server, telling him this is the random number you had sent me. So here the authentication is done only on the client side, not server side, confirming to server I did the authentication its ok. So no need for password stored on server. Is that how passkey works?
Ohh I think I got it. Instead of face recognition, one can use a passkey stored only locally not on the server, so less risk of compromise. If you use face recognition no need to use a passkey
I was worried I'd have to factory reset my tablet and phone... do you think I might have to do that or should changing the passwords and setting a passkey be sufficient?
A more secure, more convenient alternative to passwords.
Hi Leo, back in the day, when CLI was a the only thing out there, PGP for email had caught on. I used to use a product called Eudora and similar email implementations. It seems like it took a while for this to come back in a different form. Better late than never.
@@debtfordwharf It never really went away. There was always a plugin for Thunderbird, and assorted others. The problem is it's still too complex for normal people to use.
If someone adopts Passkeys, should they delete all other methods of authentication they used previously? For instance, Google Prompts? Could someone exploit/intercept Google Prompts if used at some point despite the fact that we set up Passkey?
Your explanation of pass keys is beyond excellent.
Exactly my thought to!
super clear
😮hjbfftttrr 0:33
Mk NM😅
You're so good at explaining things, it's like listening to my favorite professor.
One of the few on UA-cam that really brilliantly transmits information.
Simply incredible, makes me want to listen to you for hours
Yes, he's brilliant.
Passkeys are more secure than passwords because they are less "powerful". With an username/password pair you can potentially log in from any account, on any device, anywhere. Anyone who gets your username/password can potentially masquerade as you from anywhere. That is what makes them less secure / more dangerous. As described in the video, a given passkey is tied to a specific account and a specific device. It is important to remember that passkeys only authenticate the device and account. They do not authenticate the person. This is why the total security solution requires you keep access to the device secure. Fingerprint scanners, FaceID, or PINs HAVE to be used so that people who have physical access to your device, can't actually access your account. The device/operating system you use must, of course, provide a mechanism for securely operating a keystore.
This was an excellent video.
This is a naive point of view.
Passkeys will be stored in raw form 100% it's impossible to guarantee biometrics on everything and syncable is a core requirement for an everyday user.
Because of that you open the opportunity to steal the passkeys from a device guaranteeing that a successful attacker will have access to every. single. account. ever. forever.
As you have correctly stated, you have moved away from authenticating people to authenticating just the device. This is an inherent risk introduction.
this is the clearest, most understandable explanation of what a passkey is on youtube, or anywhere else. and it also explains that apps are actually using passkeys when they allow us to login using our fingerprint, face, or pin. and also an added bonus on what are pgp keys. im so glad i watched this. thank you very much.
Someone commented his explanation is beyond excellent - I agree and subscribed.
9:01 this is where Steve Gibson’s SQRL protocol is superior to passkeys. Both use public key encryption, but SQRL has 1 identity that creates a key pair on the fly for each login based on the site’s domain name. Elliptic curve crypto allows you to create a private key based on a determined input. The same input will always create the same key. Therefore a secret (the identity) mixed with the domain name will create a unique key pair for each login. Since this is easily calculated, there’s no need to save it for each site, just keep the original secret (identity)and recalculate based on the domain. This means the protocol and any devices can have an unlimited number of sites to log into, no extra storage and it’s easily shared between devices. Oh well, we get passkeys instead.
I guess there is no golden bullet in cryptography. Each solution has it's strengths and weaknesses, but what is considered a strength (or weakness) by one user might be the opposite for another user. Even though the passkey in it's current form is perhaps less elegant and definitely more cumbersome than the method you describe, I would still prefer the current solution. If I understand you correctly, I personally wouldn't like to have solution like SQRL that is based on a single private secret used to authenticate/unlock all (!) my accounts/logins. Sounds like the likelihood somebody would be able to crack my private secret in the future would increase with the number of accounts/logins and of course advances in technology with time.
Wow yeah, I second that. You have a very easy to understand way of explaining this!! Thanks so much. Every time I think I understand the encryption/decryption process, I seem to lose the understanding. This helped immensely.
Thank you!
Wow this was the clearest explanation I've found and finally understand!!! Thank you for this! 🙏🏻🙏🏻
This is an amazing explanation. Thank you for making it so clear.
I will be saving this video so that when anyone asks about passkeys, I will share this to them.
here we are. I thought i i already knew how, but now i really got it. Thanks from a german Guy. Great work!!!
You are a great Teacher. Especially the core puzzle is untangled with those two KEY images word by word. Great Job Leo. Thank you.
Super job Leo! Look forward to making it work...not instant.
Videos so great, speaks very well and slowly, like the words at the side, easy to understand
I consider myself a pretty savvy techie. I’ve always understood, encryption and public key and private key stuff. But for passkeys, I have some sort of mental block. That being said, Leo‘s explanation of passkeys is by far the best I have ever heard!
Agreed, & same here
I think because all other explanations contain the "and then magic happens" part, this is the first one that explains that passkeys are just ssh-like authentication with a better UI.
I've been using PGP keys for years and you have described it perfectly in this video!
Thanks!
thank you!
your explanations are so clear, far surpassing the many many other voices on this topic. Many thanks.
What i like about things like this is that they are complicated for most users and this causes things to go wrong, so you end up dropping down to passwords and email to get back in to most accounts.This negates the purpose of it. you basically bypass by clicking on the "I forgot my password" link, this mostly ends up going back to unsecured emails.
Fantastic explanation! Very clear. Been looking for a video that goes into more detail and this is exactly what I needed!
"Or have a face"
Beautifully done 😂
As ever Mr LN explains the inexplicable with ease. Been a follower for years - when internet connections required a series of morse-code-like noises and then went at speeds the common tortoise scoffed at
In a passkey-only service, isn't there a higher than normal risk of getting blocked of your own account if you lose the devices the passkey is stored on?
If your operating system drive goes down, yeah. Or if for some other reason you can't access the computer with the private keys on it.
Good question on a passkey only set up, but if you use 2 hardware passkeys (1 set as backup) do you think that solves that problem?
I see the convenience of passkeys, but for me there's still a problem. Passkeys are effectively single factor authentication. Mere possession of the passkey is generally enough to gain access to a passkey protected system. If a criminal steals your laptop and gains access to it (e.g. by shoulder-surfing your laptop password), then they can automatically access any passkey protected data you have. Using complex passwords protected by a password manager (with a strong master password) together with 2FA (using a password protected OTP generating app), whilst much less convenient, seems far more secure. In the event of a data breach at say your email provider, even if hackers got access to your email password, 2FA would still prevent them from accessing your mail.
+1 also agree with my current understanding, but I guess passkeys is still an improvement for the "below average" user that we realistically have given up on trying to make them all put in the effort to adopt the password manager + 2FA setup
No. No. No. In the description you give here, your laptop password is your first factor and the physical passkey is your second factor. That's 2FA. If you let both get stolen, it was still 2FA. If you leave your passkey in your laptop fulltime, then it's no better than using an authenticator app installed on your laptop. A passkey only becomes stronger than a laptop authenticator app if you only plug in your passkey during login and then you physically remove it and store it where it cannot be stolen simultaneously with your laptop.
A passkey is not a hardware key that you can plug and unplug from your computer. You might be thinking of a yubikey?
A passkey is a cryptographic hash that stays on your device
Good points. That's why I rely on YubiKeys for my authentication
You need a devices with fingerprint or facial authentication to create a passkey. So if your computer doesn’t have biometric authentication, you will need your phone to create and store the passkey. You’ll then confirm the access from the computer using your phone. The system is well thought and has been developed by Apple, Microsoft and Google. Nowadays all modern OS’s are compatible. The biggest problem is passkey synchronization between devices so that you don’t loose all your accesses if you loose your device. Apple has the keychain for that, Microsoft and Google will surely also have an equivalent.
Thank you! This is the best explanation of passkeys I've heard so far! One of my concerns regarding passkeys is... what happens when you have an account that's only using passkeys, have only setup passkeys for that account on a single device, and that device is lost, stolen, or is otherwise unavailable (it dies)? How do you regain access to that account? It seems like the best defense for such a situation is to have passkeys setup on multiple devices, allowing you to confirm you identity when setting up a new device after a device becomes unavailable, but that's not economically viable for some people. An alternative is to actually have a password for the service, using passkeys when possible, but that leaves the account vulnerable in the event of a data breach. Additionally, let's say I want to replace a functional device (my only device) with a new device. It seems I would need to maintain possession of that device for some "overlap period", during which I would need to login to every service I use on the new device, so that my new device can be authenticated by the old device. That seems rather cumbersome, but is probably a small price to pay for the added security of passkeys. What are your thoughts?
Each time you set up a passkey on a new machine a different form of authentication is used. For example a code to your phone, or a message to your email. Once set up it becomes your authentication mechanism. But you're always able to set it up from scratch somehow.
This is an excellent question and one that bothered me for a while. You can’t make the argument that you are in a better security position with passkeys if the use of passkeys is in addition to an authentication method that was already present. Therefore, you have only improved your security posture if you remove the old auth method and only use passkeys. However, if you do this, you run into the issue you are asking about.
I think for this scenario is exactly why having a 3rd party password manager (PM) in general, and 1Password in particular, makes sense. The PM collects and manages all the passkey private keys so no matter what happens to the device that actually created them, it doesn’t matter. You get your new phone, authenticate to the PM, and you are back in business. But now isn’t the PM vulnerable? Not with 1Password’s security architecture. There are two necessary pieces of information to access the 1P vault that are never stored in the cloud or even transmitted: your password AND a locally generated random security key. You pair those things with a hardware security key, stored in multiple secure locations, and I think you have a setup that’s nearly impossible to breach, but is also convenient
Well, I think passkeys are just a convenience mechanism in that you have to authenticate only once either in the key manager of your OS or in your password manager and then use the per device generated and stored passkeys to log in to the websites. No need to manage different passwords, and it also increases security as you are not exposing your password in your daily login routine. No chance of some man in the middle or some other malicious browser extension stealing your password.
Now your concern about a data breach happening on the website on which you use the password to login, most of the companies don't store raw passwords in their databases. They store salted one way encrypted password. As soon as you supply the password and try to log in, it is immediately encrypted in the client side and transported to the server in an SSL tunnel ("s" in HTTPS indicates that the site uses SSL, which means all traffic is encrypted)
Leo, Thanks for sharing and explaining so clearly. you are a champion!
Excellent explanation, thank you !
i needed a little extra help getting the basic premise after learning a bit about it :)
Loved that XKCD reference!
Of course! And it's so easy to remember! 😀
Best explanation on this topic. Thank you, sir!
The way u have explained is awesome thanks. I will see all ur videos
This is a much better and clearer explanation than the one provided by Google. However:
- at my last work place (a major corporation) using ssh keys was not allowed by Info Security group. According to them it violated their security standards.
- if using passkeys prevents me from accessing my Google account from a device which doesn't have my passkey, then I have to use password, which defeats the purpose
- if my passkey is automatically created on a new device, what if I'm using a public device and don't want my passkeys on it?
- with passkeys I'm relying on a strong login security of my device, but if someone breaks in into my device, then they have green light to all my accounts
1) your IT folks are misguided. :-)
2) We're moving to a passwordless future. In some cases you can choose that right now, in others you can simulate. So, no it does not defeat the purpose.
3) Passkeys will never be created without asking you first.
4) No. They'd still have to pass biometric or Windows Hello authentication when they attempt to use a passkey.
Leo,
The major concern I have regarding the password-to-passkey transition period, is that the service/company/app I am accessing will actually have both the new public key for a specific device(s) AND my original password. In the example you used where the service was hacked and they stole my public key, e-mail, etc., didn't they also get my still usable password? I mention this because I have created a few passkeys but have not seen an option to have the service permanently delete my password once the passkey was created.
Therefore, even if I create or share passkeys for all my devices to a particular service, a data breach of that service will cause the same pain it does with or without passkeys because my passwords are stored in the same old way "alongside" my public key.
What am I missing?
Thanks for the excellent video!
Karl
Microsoft allows you to remove passwords. In leiu of that, set the password to something ridiculously long and complex, and then don't save it anywhere. Nothing for hackers to steal, you'd never use it yourself, so it's as close to password-free as you can get.
@@askleonotenboom Thanks Leo. (I am in the Apple ecosystem.) Unless I misunderstood, which is entirely likely, the service still has my valid user ID and/or email and my long/complex password. Many companies use very poor data security practices and leave files exposed on cloud servers like it is a hobby, so I still believe the danger is being unable to have passwords removed from a service when you have switched entirely to passkeys. BTW, although companies frequently use poor data security practices, they are expert at apologizing after they have been breached. I currently have three different "free" identity theft services due to breaches. Being breached does not seem to cause the same reputation damage it used to.
Keep up the good work!
Karl
Have been following passkeys for a while but have never seen such a clear explanation. Congratz! Regarding passkeys i do have 2 concerns 1. Suppose i loose my device with the only private key i have, how will i be able to restore my account on a new device? 2. When creating a passkeys for an existing account, the less safe login method using a password which could be stolen from the server still exists. Hope some one can convince me that both issues can solved.
When you set up a passkey on a new device, yes, you login some other way. It could be password, but it need not be. It's more often something more secure like a confirmation email sent to the email address of record, or a text message to the phone number of record, or similar. Once you've confirmed your identity that way, the passkey is created.
Losing your device has nothing to do with any of that. ANY new device on which you want to set up a passkey goes through that process.
If you lose your device, however, once you've signed in to the account elsewhere you can remotely disable the passkey associated with that account.
Thank you so much AskLeo
The public key would be like walking down a street and writing down the house numbers you see on mailboxes, but that won't unlock the deadbolt on the front door...
So, no more passwords or password managers? What about the security of the pin? Do you need to use different pins on different devices? If I were to lose my device, wouldn’t the pin be easily compromised? Or would pins therefore need to be treated as passwords are now? My passwords are 20+ characters long & in a password manager. I question that use of passkeys & a pin is better than what I currently do.
Plus, is it easy to change the passkey if a device is changed? I assume removing a device is done by wiping the data from it. But, the old passkey for the old device needs to no longer be valid.
"no more passwords or password managers?" -- Not necessarily. People may choose to use their favorite password manager to manage their private keystore.
"What about the security of the pin?" -- The PIN is handled locally. If you have the device and you know the PIN, then you're good. Unless someone has your device, knowing the PIN alone is useless.
"Do you need to use different pins on different devices?" -- Each device is independently secured, so that would be the user's choice.
"If I were to lose my device, wouldn’t the pin be easily compromised?" -- If you lost your device, then anyone in possession of the device could access your keystore if they know your PIN. Obviously, a PIN would be easier to crack than a 20+ character password, so best not lose your device.
"I question that use of passkeys & a pin is better than what I currently do." -- Passkeys provide some significant advantages, but they have disadvantages as well. You need to choose an appropriate mix of risks that work for you. Passkeys give certain protections against phishing attacks, but are subject to being lost when the device is lost, inaccessible, or nonfunctional. Incorporating passkeys into a password manager service provides some safety here, but then there's the question of how you intend to log into your password manager.
" is it easy to change the passkey if a device is changed?" -- That will depend on the site being logged into, and whether or not you plan on managing your passkeys with a password manager service. Using a password manager provides excellent portability between devices: changing a passkey on one device would affect all other devices as well. I am confident that changing a passkey on a given site won't be a terribly difficult issue, but who knows? We still have places where your password length is limited to 16 characters (or less).
"But, the old passkey for the old device needs to no longer be valid." If you want to invalidate a passkey (if it's tied to a device, instead of a password manager), that would be done on a site-by-site basis.
@@neuideas thank you very much for your clear answers to my questions!! A big help!
@@neuideas John Cole One query
If a passkey can be uploaded on a online password manager and can be used to login on a new device, then How come is it different from a password mechanism ??
Also, accessing password manager using Local pin, fingerprint or face lock is not possible since its a new device ??
I still cannot understand how I will login on a new device if its lost if I own only a single device
@@himanshuchhabra1942 " is it different from a password mechanism ??"
It's different because it uses public key verification of digital signing, rather than hashing a password. The burden shifts from remembering your login information, to not losing your private keys. The private keys are never revealed, so they are never risked.
"Also, accessing password manager using Local pin, fingerprint or face lock is not possible since its a new device ??"
You will need to authenticate the application first. This can be done by validating your account on an already authenticated device.
"I still cannot understand how I will login on a new device if its lost if I own only a single device"
You need to set up a backup device first, before you lose your primary device. Alternately, you may be able to log in using one-time use backup codes, if they are offered. Beyond that, you could always fall back on standard login credentials, assuming that's an option.
@@neuideas
I understood the mechanism , I was trying to understand the big picture. My point was private keys revelation is not needed, if the hacker can use other ways of authentication which are provided for the user in case he loses the device.
Leo, this was an absolutely superb presentation. I already had a decent understanding of private/public key pair but I knew nothing about passkey. I'm going to see how my very non-technical wife can follow your video now :-)
Great explanation. From a privacy perspective, I’ve always wondered about adding a fingerprint or face as a ‘passkey’. Is this stored anywhere that compromises one’s privacy?
I don't believe so. It's stored and used only on the device.
Thanks for your nice overview Leo. I’m interested in how third party password manager apps will help manage this information, versus the device operating system itself.
3rd party password apps aren't necessary. The key pairs are known only to your computer (the private key) and the site you're accessing (the public key). No password for 3rd party password apps to manage.
@@HarshColby Some online 3rd party password managers can store your private key. That's how they sync your passkey between all devices where the password manager works. Some of those online apps include 1Password and Bitwarden. KeepassXC is an offline password manager that will soon support passkeys, too.
Thank you. You described a software pair of keys. And Google does supply that. But, there are vendors selling hardware devices. I assume that using hardware PassKeys, the public key is identical on each of the web services that I use the hardware PassKey. When should I consider buying the hardware PassKey? Do most web services also require a password in addition to the PassKey?
I haven't seen hardware passkeys. Please don't confuse Yubikey devices with this. They are two-factor keys.
Actually if the yubikey can provide user verification via a pin, fingerprint, or something like that, it’s considered a passkey! Most people use them for 2SV, but they’re very much usable for passkeys (given they can perform user verification)
@@askleonotenboom Yubikey 5 series can be used for passkey (FIDO2/Webauthn) authentication. Actually, that seems to be the only (simple) way currently to use passkey authentication on Linux (not Android) devices.
Outstanding - I subscribed immediately
Excellent discussion of the theory. Clarified a lot of questions I had.
Thank you for your thorough overview. Cheers!
Impressive video. If anything, the private key is the weak link. So I am left with the doubt that the private key is safe. I know that if one has possession of the hardware with linux the logon procedure is not going to be much of protection. The only protection I would trust in that case is both disk encryption and a logon password. And make sure to switch your computer off, or someone might add a password to the list of passwords for disk encryption (LUKS). However, even that has a shelf life as quantum computing is around the corner. Of course the whole encryption scene will change by that time. Anyway, I thank you for your explanation sir, very clear!
If I create the passkey on my device (smartphone) and a private key is generated based on data from my device and my biometric data, the only way to compromise a passkey-protected account would be to hack my password manager ? Or did I not understand well? The big difference is that I, the user, do not know my private key as I do for the password as it is a very long and complex alphanumeric string.
Great video Leo. I can't wait for passkeys to take over the password phase. Do you have a list of services that have already started using passkeys, besides google?
www.passkeys.com/whos-using-it
Gee, I finally know about passkeys. I was so curious about them. One problem, however: if someone breaks into your house, and if you are not there and your computer is turned on, they can just sit down at your computer and login anywhere, can they not? Maybe the operating system would ask them for a pin, or a fingerprint.....
Exactly. The OS will ask for that.
@@askleonotenboom Is this 2FA? Exactly how?
@@viktorpaulsen627 Not really, no. It's closer to a plain old password replacement that's more secure. Kinda. Some think of it as 2FA because your device will prompt you for your PIN/fingerprint/face before providing a passkey, but that's still only one factor that you had to provide in the moment.
Great update Leo, Passkeys for Google Accounts are now available.
Pretty sure I mentioned that in the video, or at least the companion article. :-)
Hmmm, thanks for the great explanation . I wish google would hire you to explain. Dreadful articles. However I am already deep in the swamp of questions. Such as- your laptop has problems any you have to give it to a repair person (not a hard drive replacement). They need admin rights. Does that leave you open? Also, another question - in an emergency (you can’t give them data) how do they get into your device to pull data? Lots and lots of questions before the complicated users feel comfortable. Again, thanks
"Does it leave you open" yes. Choose trustworthy repair people. (Or take extra steps to secure sensitive data while still leaving the machine operable.) Not understanding the emergency scenario you're describing, though.
Make sure you always make backups of your devices and authenticate with multiple devices if you use something like passkey
Just found you. Outstanding succinct explanation thank You. Subbed of course
Closet somebody reminds of Bob Ross, soothing and mesmerizing. And as a bonus you get educated.
What happens if the machine with the passkey dies? How would you be able to access the account. I'm thinking here of things like cloud storage.
Each machine has it's own passkey. So you'd be starting over as outlined in the video/article by signing in some other way.
@@askleonotenboom But if it's possible to sign in some other way doesn't that lower the level of security? I apologize for seeming dense but considering the skills of those with less than honorable intentions and the amount of information stored in the cloud this system may be an improvement on the current model but it still isn't perfect. I suppose anyone with these concerns could just have two machines with access then if one died you wouldn't be stuck. It will be interesting to see how it will all work out in the end.
Thanks for your prompt reply.
So, it should be strongly recommended to generate passkeys from more than one device for each account, inmediately after creation or activation of passkeys.
@@mfr2 Not necessarily. It depends on the service, but like I said, you probably signed up with an email address so an email to that address could also confirm you're you.
@@frederickclause2694 Of course it's not perfect. There's no such thing. But it is significantly more secure than password based authentication. AND it's easier to use. 🙂
As a senior citizen I am very concerned that when something happens to me, my children can access all my accounts and information. If I set up a passkey, do they have to have my device to do so?? Right now, I keep passwords in an encrypted file for which they have the password and I send them the current file on a semi-regular basis.
This sounds wonderful if YOU are the only one using your device and the only one needing to log into accounts. But I am a bit confused about how it will work in a situation where multiple people need to access the same account (a bank is an example).
Whatever technique the service uses for you to set up a passkey on a new machine should be made available to your heirs.
Where does PGP email encryption, decryption, digitally signing emails fall in the mix? Sorry, I am not a techy, just someone trying to learn and understand. Thanks.
Related encryption technology, but used for a different purpose.
Hello.Thank you for this very interesting and informative video. A question of security: I imagine that the public key and the private key are created using an algorithm that ensures the link between the 2. What happens if a hacker gains access to this algorithm? Can he decrypt the private key? This is a very unlikely hypothesis and the risk of ordinary passwords is certainly much greater.
Now that TSA and flight security systems around the globe and immigrations check points are using fingerprint and facial identity, what could possibly go wrong? I would think that a pin number would be a better choice for the final authentication, while using these passkeys.
This is an absolutely fantastic video - thank you!
What if I have only one device storing my private key (say my phone) and I lost it. What is the recovery option here? and the person who was able to get my lost phone figured out the login pin? I am not questioning the security of passkey but trying to understand this scenario. In case of password, I know it and can use it from any device. By the way excellent explanation!
There is ALWAYS another way to sign in. That other way may involve more steps and be less convenient (say, emailing you a code), but think about how you establish a passkey to begin with: you have to login somehow. Once you're logged in you can then revoke the passkey assigned to your phone.
If the site owner isnt forced to update to passkeys it will take ages for this to be implemented.
And if its mor conveinient can be discussed. I have setup passkey on my google account. So everytime i log in i have input my win passcode. Compaired to chrome just remembering the password. Not easier. Or i have to buy a camera or a fingerprint reader for my desktop.
And how do i know if my fingerprint/face dont get stolen by hackers exploiting flaws in the camera/fingerprint reader?
Imo just a new set problems compaired to the old way.
Does the rollout of pass keys present a vulnerability for accounts that have not been set up yet that may be compromised, allowing a threat actor to set up the pass key on another device? Does the key have a method of being reestablished manually or during a password reset or account recovery process?
A passkey requires that you be able to authenticate some other way in order to be set up. A devices passkey can be individually revoked remotely.
Subscribed! Thanks!
Excellent & clearly spelled out...thanks!
thnx. one question, what i missed explaind. say you have done so for your phone. all is working fine. you now get a new phone. do you have to start all over again on your new phone for every account (as where it a other device as you mentioned) , and then can without a problem get access to that account ? just wondering....thnx for the feedback
Depends on the system you're signing into. Worst case you start over, but in general it could be as simple as a one-time additional hoop to jump through (a text message to confirm, and email to respond to, or another device on which to approve the sign in).
You explained in a detailed way. Passkey is still in infacy stage, I'm still waiting another 1-2 yrs..
It would be nice if the user could decide which data has to be entered each time they log in. I would use username, password, 2FA and passkey together. The passkey would protect against an attacker being able to log in with my remaining login data from someone else's computer.
The point of potential confusion is where the private key is going to be stored. It’s not obvious.
If you live in a multi platform environment where you login to website from various OS and mobile devices running different OSes, who is the private key store cross platform?
Do I use my Google account? Do I use my 3rd party password manager? My 2 or 3 OSes?
I’m already concerned that Google Authenticator offers to backup your 2FA keys in the cloud, thats putting all your access keys in one basket and outside of your physical control. I only use the QR export and import to a separate device not via the cloud.
If I put passkeys in 1Password, what if my account is hacked? All the eggs are in one basket.
PKI relies on the private key not getting duplicated outside of your control. This is the weakest link if we allow the private keys out into the cloud.
Each platform gets its own, unique, private key the first time you sign in, and it's stored in that platforms secure store.
Tools like 1Password can act as that store to make passkeys available across platforms if you like.
@@askleonotenboom "In theory".
As I've been testing passkeys cross platform (including Yubikey, Google and Microsoft Authenticators), I've yet to get the 1password passkey for my Adobe account (as a a test) to work x-platform. My android 14 phone and iPad OS 17 aren't accessing 1password for the passkey.
The xp functionality still needs a lot of work.
My question is, what happens if you create a passkey and use it on your mobile device for your crypto account. Is there any instance where you could lose any kind of access to your passkey, to where you couldn't access your crypto account or mobile device?
Thanks. That was top notch.
I saw this on my PlayStation account I have lost accounts before so I was looking for a different way to keep my account safe but I didn't understand it thank you for this video
For how long will the private key be valid ? Does it expire ? (Like certificates do)
Greetings,
Rik
To the best of my knowledge they do not expire.
Excellent expositor! Thank you very very much.
Okay, I know and understand that passkeys > passwords, but I still have a few questions. Your house burns down and with it your laptop, desktop, tablet, and phone. How do you recover those passkeys? Now I understand how this can be done, but so will the black hat hackers in Elbownia. What can we do to prevent that attack?
Next, what about brute force on the public key? Yeah, it would take a zillion years to do it on a laptop today. We had the same thing with the Windows NT password system back in the 90s, and just a few (like 4-5) years laters, it was possible to crack within hours using a desktop computer. With all the computing power on earth today (equal to a single GPU in 2050), how long does it take to get through all the possible combinations? Are there collisions (you create an A-B pair, but it turns out that R can also Decrypt some A?
Can there be a 'dictionary' of passkeys? There are a number of formulas and prime numbers in use. Once they are fed through, is it possible to recognize the A key and then pull the B key out of a dictionary as well?
Regarding the house burning down question. If you're in the Apple ecosystem, your passkeys are stored on your iCloud Keychain. If all of your Apple devices are destroyed in the fire, when you go to set up a replacement device you'll need to go through their Apple account recovery process (assuming you lost the appleid/password in the fire too). It's a good idea to prepare for this beforehand (for example, you can nominate someone else's account as your recovery buddy). I believe Google has a similar password manager these days.
wonderfully explain. Thanks, Sir.💟💟🎀🎀
Great explanation and examples, but I have a question! Let say that we are in 2027 all my accounts are using passkeys and there are no passwords anymore and all the private keys are stored in my machine. Now, let's imagine that someone stole my machine or it gets damaged, how can I access my accounts again without my private keys?
Your passkey is never the only way authenticate. Remember, in order to set up the passkey on each device you had to authenticate some way that didn't involve a passkey. So that continues to work when you move to a new device. A common password-free example is a link sent to your email account.
They must have "both public and private key codes" in the file or operating system. Otherwise system cannot proceed with the matching process. Hackers can find the way steal the private key code. By far the best method is to send a temporary code to your private device. Use it as your private code.
It's a split key pair: one key is on (only) the server and the other is on (only) your device.
This is a superb explanation. I have one dumb question: is there no way for the server side and the device side to be hacked at the same time, or even at different times, and somehow allow the private and public keys to be mated up by the hackers?
Seems incredibly unlikely. (As in infinitesimal chance.)
Hay Leo can I use the same passkey on different devices? Or do I need more passkey Thanks
Some password managers like 1Password will store your passkey for use elsewhere. Otherwise, passkeys are per-device, so you'd set one up on each device the first time you sign in to the account on each device.
Can you use emojis in passwords? It would sure expand options
It depends entirely on the system asking you for a password.
I wonder if they hijack but knowing hardware locked to the key. But what if the emulate the hardware?
What happens if you hard drive crashes? Will the passkey be on the backup. What if you have no backup. How do you get in the account?
The same way you got into the account when you first setup a passkey on that machine. Usually a different, more cumbersome, authentication process, like an email to your email address with a code or something similar.
Let's say passkey is the only way to login to an account. If you have it on one device only, then losing the device = loosing account. Also if you just happen to go on a trip and forget that device with the passkey then you can't login. I sort of prefer to not be blocked from my account like that
Passkeys are NEVER the only way to sign in to an account. You had to sign in some other way in order to setup the passkey in the first palce.
Good clear explanation. My concern is that each device is locked to each passkey. This potentially could be a lot to manage. Also a source of privacy issues. Given the exponential growth of surveillance capitalism I will avoid use of passkey until it's better addressed.
Not much to manage. Registering a device requires a click and is thereafter transparent. Apple and Google😢 offer the option to store the private keys in the cloud so that one key-pair can be used across multiple devices. The choice depends in part upon which you trust more, your device or the cloud.
My understanding is that if I want to use another device to check my email, it had to be authenticated. Each key pair uses the device ID as part of the encryption. Not normally a problem, but if I want to use my friends or public computer that might be an issue.
Very well explained!
Passkey, it is the server provide a challenge to the user to be signed by private key or server signs a challenge and ask the client what the challenge is?
In the context of Passkeys, the server typically provides a challenge to the user (or client) that must be signed using the user's private key. Here’s a brief overview of the process:
Challenge Generation: The server generates a unique, random challenge (a nonce) and sends it to the client (user).
Signing the Challenge: The client uses their private key to sign the challenge. This ensures that the client possesses the private key without revealing it.
Response to the Server: The client sends the signed challenge back to the server.
Verification: The server verifies the signed challenge using the public key associated with the user's account. If the signature is valid, it confirms that the user has successfully authenticated.
My main concern about this is that I can't easily imagine what could be a smooth transition to a new device (computer or phone for example) when the previous one is damaged or simply obsolete.
It's exactly the same as setting up the passkey on that first device the first time. You authenticate some other way, and set it up again.
@@askleonotenboomThe first time it's one site at a time and progressively.
When switching computer of phone with hundreds of accounts already registered with the previous device (I've more than 800 entries in the "personal" section of my KeePass database) it could be not so trivial.
superb explanation of passkey done layman terms.
The source of “correct horse battery staple” is an XKCD cartoon. Number 936 to be exact.
I'm trying to figure out what happens if you lose access to the device which authenticates the credentials.... like the fingerprint sensor on my phone ... if my phone explodes or gets stolen... how am I going to gain access to my account?
The same way you got access to it when you set up your passkey in the first place. There's ALWAYS another, often less convenient, way to sign in.
It's a good explanation, but I am going to create a fictitious email account and set up a passkey with a fingerprint. Then I will try to access the fake account via a desktop that doesn't have a fingerprint scanner. This is where I have trouble seeing how the same account can be accessed by a phone, and desktop and a Chromebook without f*rting around with all sorts of different pins.
Thank you for that explanation.
I may be a novice on this topic, but I thought that authentication with a private/public keypair was a matter of signing a "nonce" (a randomly generated number) with your private key, and the server using the public key to verify the signature. Am I wrong?
I believe you're correct. What you're referring to as a nonce is essentially the "I'm thinking of a number" referenced in the video. I did say it's a very high level overview and abstraction. 😀
TY 8) What if you loose your PC with private keys or have the disk died?
Then you'll use a different recovery method to login on a new machine.
Hello. What if i loose device? Can i still access my data and how? Thanks
Of course. You had to setup the passkey on that device by authenticating some other way to begin with, so you would simply authenticate some other way on a new device and setup a new passkey.
Excellent explanation - thanks
Let me see if I understand. With password, both clients and server store the pre agreed password. The exchange between the two is done by creating a secure channel with public and private keys, and then I send the password from client to server to get authenticated. With passkey, we still create a secure channel using public and private keys, but this time the server sends a challenge (say random number) to client. Client authenticates (say face recognition) authentication done on client side. Then it sends the challenge back to server, telling him this is the random number you had sent me. So here the authentication is done only on the client side, not server side, confirming to server I did the authentication its ok. So no need for password stored on server. Is that how passkey works?
But then I don’t understand what is the passkey that still needs to be used?
Ohh I think I got it. Instead of face recognition, one can use a passkey stored only locally not on the server, so less risk of compromise. If you use face recognition no need to use a passkey
Finally, I understand!
How do you sign out of a machine so that if someone else signs into it after you does not have access to your key?
They have to sign in to a different account on that machine. Passkeys are per user.
Interesting. I am familiar with PGP, for example, but had not realised it could be used in this way.
It's one of those technologies I find endlessly fascinating.
Very good explanation
If someone has my passwrds already and has access to everything will this work to stop them getting into my phone?
Change your passwords first.
Oh wow that's a really quick response, thank you sir.
Because this person has my email will they not see that?
I was worried I'd have to factory reset my tablet and phone... do you think I might have to do that or should changing the passwords and setting a passkey be sufficient?