The latent limitation of artificial intelligence

Multi-format AI bias workshop series

Click to load imageA screenshot from a Stable Diffusion UI, where the prompt is a photo of five oranges. None of the images below have five oranges.

To reduce website CO2e emissions, project images need to be manually loaded.

A screenshot from a Stable Diffusion UI, where the prompt is a photo of five oranges. None of the images below have five oranges.

Click to load imageTwo screenshots from Stable Diffusion UIs. One prompt is a photo of a man, the other is a photo of a woman. The resulting photos look quite similar.

To reduce website CO2e emissions, project images need to be manually loaded.

Two screenshots from Stable Diffusion UIs. One prompt is a photo of a man, the other is a photo of a woman. The resulting photos look quite similar.

Click to load imageBlack text on a white background with a 6-step guide on how to explore AI chatbot search errors.

To reduce website CO2e emissions, project images need to be manually loaded.

Black text on a white background with a 6-step guide on how to explore AI chatbot search errors.

Click to load imageThe top of a research article called ChatGPT is bullshit.

To reduce website CO2e emissions, project images need to be manually loaded.

The top of a research article called ChatGPT is bullshit.

About

The latent limitation of artificial intelligence is a multi-format workshop series about AI bias, bugs and bullshit.

The series revolves around the central theme of investigating AI through hands-on demonstrations and experiments, but its specific focus changes from workshop to workshop. So far, the series has covered AI image generator biases for architects, large-language model bullshit for queer people, AI chatbot search bugs for researchers and a handful of other topics.

The workshop series has so far been conducted in two different ways. First, as a "traditional" workshop, where participants are given a brief presentation before asked to complete certain experiments by using AI themselves. The other workshop format is a mix between a performance and workshop, where I go through tasks myself on stage, but include the audience in various ways during the performance. The first type works best with fewer people, whereas the second one can be scaled up to larger audiences.

The goal of the workshop series is to pull back the curtain on AI and draw attention to the biases, bugs and bullshit inherent to the technology, which in turn enables the audience to better understand what AI can and can't do.

Events

CO2e emitted: ~0.00g