Do you work in AI? Are you tired of people using images of the terminator, robots, or glowing brains when they use your work? Do they make you uncomfortable but are you having trouble articulating why?

I have been there.

In this talk I will introduce you to the *"Better Images of AI"** project. It consists of an image bank of better images, and a guide that explains why those tropes are unhelpful and what you should look for in alternative images.

As a bonus, I will also share how I created the images I have contributed to the image bank, using not Stable Diffusion, but good ol' Scikit-Learn!

At the end of this talk you will be able to discuss with your peers what's wrong with popular images of AI , where to find better images of AI, and even have some inspiration on how to create your own better images of AI.

rens dimmendaal

Affiliation: Xebia

Rens has been playing with generative AI for years. Whether that means rhyming with BERT, or creating "art" with Scikit-Learn...or now with the latest generation of GenAI models.

At work he is a Principal Data Scientist at Xebia Data. There he leads teams of Data Scientists and Machine Learning Engineers, focusing on delivering value through AI-powered products and MLOps best practices. Rens shares his knowledge and experience at meetups and conferences, such as PyData, MLOps.Community, and NeurIPS.

visit the speaker at: GithubHomepage