speech

AI-generated content surrounding twitter medical Q&A

Inspiration

Medical knowledge is expanding. Keeping up is challenging. Learning opportunities are diverse but demand a huge amount of manual work. AI-generated content can increase speed and reduce effort to create content.

What it does

This project offers additional content ("tell me more") surrounding the frequent Q&A sessions of the @Knowmedge twitter channel. The additional content can be generated on-the-fly and requires supervision only to check for correctness.

How we built it

Main part is the parsing of topics (i.e. tweets) and generating content using the cohere.ai API. In addition a streamlit frontend is provided.

Challenges we ran into

AI-generated content needs to be check for correctness even though the effort is much less compared to generating content manually.

Accomplishments that we're proud of

Accomplishing a pipeline of feeding topics to the provided API and displaying the generated content using a fit-for-purpose frontend.

What we learned

We got an impression of the possibilities of AI-generated content and huge language models. We learned about the parameters of temperature, tokens and presence penalty. Furthermore additional app development and deployment skills have been applied.