New Feature Demo: Configurable Safety

Поділитися
Вставка
  • Опубліковано 12 лип 2023
  • At Inworld, we acknowledge the importance of content regulation and player safety. We have developed a robust Configurable Safety feature for our platform, providing developers with control over the topics their characters can discuss.
    Our safety controls are designed to ensure that AI NPCs speak appropriately within their game universe, irrespective of whether the game is rated "E for Everyone" or "M for Mature." Developers can adjust game dialogue around topics such as profanity, violence, mature themes, alcohol, substance use, politics, and religion. However, our system prohibits any form of hate speech or content encouraging self-harm.
    These safety measures aim to enhance the gaming experience for all players, while ensuring characters remain within their established roles, comply with international laws, and avoid interactions that could harm your company or game franchise's reputation.
    🧠 Create AI characters for free: bit.ly/4akcuKU
    🕹 Learn about AI-powered gameplay: bit.ly/49mSJ3V
    🛠 Build with dev docs: bit.ly/3VFNeKF
    🎮 Join our Discord - talk AI and gaming: / discord
    #gaming #gamedev #gamedevelopment #games #gaming #indiegames #artificialintelligence #generativeart #aicharacters #chatbot #chatbotdevelopment #unrealengine #unity #unitygameengine #oculus #metaverse #avatar #virtualreality #madeinunreal #virtualcharacter #madeinunity #gpt3 #naturallanguageprocessing #chatgpt #chatgpt #globalgamejam #npcsarebecomingsmart #npcs #AINPC #NPCAI #chatGPT
  • Ігри

КОМЕНТАРІ • 14

  • @kiyasuihito
    @kiyasuihito 5 місяців тому +1

    This project is seriously mind blowing. 🎉 I wish there was a way to host the models locally though.

  • @Retrofordguy
    @Retrofordguy 10 місяців тому +4

    Ive been messing around with some characters and ive noticed it breaks immersion pretty fast when something with even the slightest wiff of explicit and it goes into shutting you down. Now as somone who tried it as a therapy device some things need to be said otherwise you end up with lost immersion and end up giving up.
    I understand its difficult to police but its somtimes important to have certain convos without it immediately shutting you down

  • @gordy12gg
    @gordy12gg 10 місяців тому +5

    Will this be on regular accounts too?

    • @inworldai
      @inworldai  9 місяців тому +1

      We have limited the feature to Enterprise plans at the moment as we build it out and improve it.

    • @gordy12gg
      @gordy12gg 9 місяців тому +1

      @@inworldai ok!!

  • @gregstuck3661
    @gregstuck3661 10 місяців тому +2

    I have created a few characters and having fun experimenting with the various settings - however I cannot find the settings tab for "Configurable Safety". Can someone please explain where these settings are to be found? Thx.

    • @inworldai
      @inworldai  9 місяців тому +2

      Currently, this feature is only available in our Enterprise plan as we work to improve and scale it.

  • @cwishno
    @cwishno 10 місяців тому +1

    Is the Configurable Safety implemented yet on free accounts? I tried to find the Configurable Safety feature but cant find it.

    • @inworldai
      @inworldai  9 місяців тому +1

      It's currently only available on Enterprise accounts as we build out and improve the feature.

  • @gordy12gg
    @gordy12gg 10 місяців тому +2

    Cool!!

  • @schnek8927
    @schnek8927 10 місяців тому +8

    0:45 There is no such thing as ”hate speech”.
    This is especially true for the creation of AI characters. What introduced me to inworld was a video of it being implemented into Skyrim, and in such a world ”hate speech” (being openly critical of other groups) adds to it’s realism as, well, Skyrim.
    Nothing is more unimmersive and unfitting than a character suddenly going on political rants about their fourth-wall-breaking ideology.
    At the VERY least, disable the lecturing. It should NEVER speak like a soulless moral busybody.
    Ideally, it would react appropriately to the character, but i guess that’s not a priority.

    • @DiveTheseClips
      @DiveTheseClips 10 місяців тому +1

      Yep. There are many fictional universes with lost of hate in them. Quite often it is the main story driver - fighting against hate and injustice.
      But if all the AI bad guys will end up lobotomised like chat gpt then all the development effort will be for nothing.

    • @khairulhamdan762
      @khairulhamdan762 10 місяців тому

      Not only that but in their own demo they have anti-robot hate speech... that is a little bit hypocritical don't you think @inworldai ?

  • @lukasmerten7697
    @lukasmerten7697 Місяць тому

    How can you be proud of never allowing fictual characters to commit hate speech? Do you not trust the average consumer to judge and interpret fictional game characters? What if I wanted to make a game about fighting a racist autocratic regime with a racist leader on top, and when I interact with him he cannot say anything racist!? Plus I can never make a villain character that hates all women because of whatever trauma his wife caused him before? Even in nationality, I cannot create a character that deeply hates all people from an enemy nation?
    That's bullshit. I hope I misunderstood this...