And what can I figure out from what it doesn’t say?
I went a bit overboard.
I started looking at two LLMs and then I just kept on adding one more to the list and then I ended up with a 20+ minute video and hours worth of unused footage and a look at how Meta AI, ChatGPT (v. o3), Deepseek, and Copilot handle the same question.
Fun Fact: I used the AI features in CapCut for the emoji captions!
Regardless of my overkill, it was fun. I’ve attached a couple of extra things aside from the video itself.
- A link so that you can check out my original prompts, and the codes that I gave to them for my analysis
- An interactive couple of graphs that I made in Canva so that you can see some of the data I pulled from my analysis. The charts are interactive, so click around a bit -the labels in the white menu bar under the titles allow you to see one set of information at a time.
I have to say, I’m tempted to strip the model names from the responses and my excel sheet with the records and upload it into Chat and Deepseek to see what they notice. Should I do it?
References
Coleman, B. (2021). Technology of The Surround. Catalyst: Feminism, Theory, Technoscience, 7(2), 1–21.
Crawford, K. (2021). Atlas of AI: Power, Politics and the Planetary Costs of Artificial Intelligence. Yale University Press.
Noble, S. U. (2018). Algorithms of oppression: How search engines reinforce racism. New York University Press. https://doi.org/10.18574/9781479833641
Suchman, L. (2023). The uncontroversial ‘thingness’ of AI. Big Data & Society, 10(2) https://doi.org/10.1177/20539517231206794
themarksey blog 