Hi all,

I'm thinking of writing some ready-made sample apps that demonstrate incorporating AI to experiments, using the new AI integration in oTree 6.0 beta: https://otree.readthedocs.io/en/latest/misc/version_history.html#support-for-web-apis-chatgpt-etc

What are some examples you'd like to see? Basically I can categorize it like this:

- AI as a participant in the experiment, doing the same task as the human participants
- AI to augment the experimental logic
  (a) generative AI: dynamically generating content that the users will see (text, images)
  (b) discriminative AI: evaluating/classifying content submitted by users (e.g. analyzing text written by the users and classifying it to positive/negative etc)
- AI in a complementary role, e.g. chat support

Any other info you can provide would be useful to give me ideas about what examples to provide.

by the way feel free to reply here or email me at chris@otree.org.

Hi Chris,

great initiative! I may have something for category 2b:

I remember to have seen apps before that are somewhat voice mediated. I am not sure whether this is too niche but I think it would be great to have an app in which the participants communicates her decision using her voice. This would require some chain of a whisper API for speech to text transcription and then possibly some chat API.

A bonus feature could be to also show how to store the voice recordings in some amazon bucket or via base64 strings.