Exploring the Depths of Debate with ChatGPT: A New Frontier

“Moriarty's creation was unintentional and the direct result of a discussion between Data, La Forge, and Doctor Katherine Pulaski about Data's ability to solve problems like Sherlock Holmes... La Forge… decided to let the computer create an opponent for Data, and told it to create an adversary with the ability to defeat Data, not just Sherlock Holmes.” -James Moriarty (hologram), Memory Alpha

I challenged Chat GPT to a livestream debate. I argued for positions I don’t hold—in favor of diversity, equity, and inclusion (DEI), while ChatGPT presented counterarguments. The result was surprisingly engaging and it highlighted some fundamental differences in first principles.

The DEI Debate

In our debate, I championed DEI as an intrinsic and overriding good, emphasizing the importance of having historically oppressed groups in positions of authority. I argued that placing minorities in unearned positions was more important than traditional metrics of competence or meritocracy. ChatGPT assumed, without argument, that competence hierarchies and the primary missions of organizations (such as hospitals healing people or fire departments putting out fires) overrode diversity considerations.

This clash of first principles made the conversation feel like "two ships sailing past each other in the night." ChatGPT never directly argued against my foundational belief that DEI is an overriding good, instead it focused on merit, competence, and distrust of the majority toward the minority. This highlighted a significant gap in how it approached the debate, that is, accepting by fiat its bedrock assumptions instead of reasoning to them.

The Self-Defense Debate

Midway through the livestream, we switched gears. I had ChatGPT argue as if it were Peter Boghossian, and we debated the importance of self-defense. ChatGPT took the stance that self-defense is crucial, while I argued against it. Unfortunately, we were cut off due to ChatGPT’s technical limitations, but I found the discussion up to that point quite engaging. ChatGPT's arguments were straightforward and displayed a remarkable common sense, it made no radical or extraordinary claims, but it presented the claims clearly.

Reflections And Future Experiments