Amazing... In the previous video when you asked about the uses of Chat GPT with storyline this was exactly what I was thinking of... I imagine it is very hard to do, but it as an amazing result
AI is a game-changer in our business,too. Next step could be to transform the text from the AI to speech. Would love to see the JS script which connects SL to ChatGPT 🙂
A very valid comment and it is something we want to investigate also. For example, if we were to design our prompt in order to capture those edge cases, and how effective the 4.0 model is for recognising those situations and returning a valid response. Those companies like Duolingo etc. that are using 4.0 in production right now must have found a way to manage invalid user submissions, and they are using the exact same 4.0 model as is currently being rolled out to other developers, so there must be a way of defining those tolerances to a certain degree...
Is it possible ChatGPT could not give the required response. For example if I write something inappropriate would it still respond with comet's or would it say..as an AI language model...or I don't have access to the internet or my cut off period is 2021...
That's a great question! I hadn't considered how it would respond to questions which ChatGPT would tend to avoid like whether it believes in something... In this case attempting to ask about something like who won the Premier League in 2022, or if it believed in ghosts, it would still formulate its response based on my request for what the user had typed to be analysed against our three company values, so we get back are interesting responses like "Your love of sports is admirable, but at Discover eLearning we instead value X, Y and Z..."
Magnificent- do you have details you could share about how this is sent to chatgpt? Conceptually, I understand everything noted here, but like many in our field, it is that "connective tissue" between systems that is the failpoint.
Thanks David. We are generating the cURL POST requests locally at this moment during testing. Our published Storyline application authenticates itself and submits each request to the gpt-3.5-turbo API through a separate application we have built running on a local server.
Absolutely I believe this can be done, and is something I am experimenting with right now! I will post some videos of the results as soon as I have something cool to demonstrate. CH
How to do this? Can you make some tutorial? I'm struggling with that from few days. I making somewhere a mistake (I think somewhere in javascript) but I don't know where. It can also be paid.
We have a full Advanced eLearning Developer Series course coming very soon! You can register your interest here: discoverelearninguk.com/store/create-elearning-using-chatgpt-and-articulate-storyline-360/
This is amazing work, do you have a guide on how we could implement this, I would love to play about with it for a portfolio project I want to start for building soft skills. From what I took from your video, this could hopefully work well in a conversational mode to help contact centre advisors get valuable feedback on certain scenarios of soft skills and conversation building
can you explain how you entered all these settings in both ChatGPT and then how you built it inside Storyline? That would be so helpful. THank you
Amazing... In the previous video when you asked about the uses of Chat GPT with storyline this was exactly what I was thinking of... I imagine it is very hard to do, but it as an amazing result
Could you share the steps on how to make this happen?
Same here. Knowing how you did it step-by-step will be an amazing tutorial!
AI is a game-changer in our business,too. Next step could be to transform the text from the AI to speech.
Would love to see the JS script which connects SL to ChatGPT 🙂
I agree. I would love to see the JS code
Very good use of Chat GTP my only concern atm is how it would respond to a totally random answer. It's in the testing I suppose. Thanks for sharing.
A very valid comment and it is something we want to investigate also. For example, if we were to design our prompt in order to capture those edge cases, and how effective the 4.0 model is for recognising those situations and returning a valid response. Those companies like Duolingo etc. that are using 4.0 in production right now must have found a way to manage invalid user submissions, and they are using the exact same 4.0 model as is currently being rolled out to other developers, so there must be a way of defining those tolerances to a certain degree...
Is it possible ChatGPT could not give the required response. For example if I write something inappropriate would it still respond with comet's or would it say..as an AI language model...or I don't have access to the internet or my cut off period is 2021...
That's a great question! I hadn't considered how it would respond to questions which ChatGPT would tend to avoid like whether it believes in something... In this case attempting to ask about something like who won the Premier League in 2022, or if it believed in ghosts, it would still formulate its response based on my request for what the user had typed to be analysed against our three company values, so we get back are interesting responses like "Your love of sports is admirable, but at Discover eLearning we instead value X, Y and Z..."
Magnificent- do you have details you could share about how this is sent to chatgpt?
Conceptually, I understand everything noted here, but like many in our field, it is that "connective tissue" between systems that is the failpoint.
Thanks David. We are generating the cURL POST requests locally at this moment during testing. Our published Storyline application authenticates itself and submits each request to the gpt-3.5-turbo API through a separate application we have built running on a local server.
Do you think if it's possible to build a conversation between us human with the GPT API along with the text to speech api built in using Articulate?
Absolutely I believe this can be done, and is something I am experimenting with right now! I will post some videos of the results as soon as I have something cool to demonstrate. CH
@@discoverelearning I just purchased your new course earlier and hopefully it will be included in the GPT course too.
It will be great if you can show us how to do it.
How to do this? Can you make some tutorial? I'm struggling with that from few days. I making somewhere a mistake (I think somewhere in javascript) but I don't know where. It can also be paid.
We have a full Advanced eLearning Developer Series course coming very soon! You can register your interest here: discoverelearninguk.com/store/create-elearning-using-chatgpt-and-articulate-storyline-360/
@@discoverelearning Done! 🙂
This is amazing work, do you have a guide on how we could implement this, I would love to play about with it for a portfolio project I want to start for building soft skills. From what I took from your video, this could hopefully work well in a conversational mode to help contact centre advisors get valuable feedback on certain scenarios of soft skills and conversation building