Why ChatGPT is wrong and artificial intelligence is dumb and dangerous

Last year, I did a talk about artificial intelligence. A talk on why artificial intelligence is not that intelligent at all. This session is a follow up.

I will give specific examples on what can go wrong dangerously trusting tools like ChatGPT. A lot of people interested in AI and ChatGPT in general, are talking in technical terms, like LLM. However, in the AI field, other terms like “hallucination” and “anthropomorphizing” are real phenomena as well.

A lot of times, we attribute human characteristics to AI, which leads to interpreting its results wrong. As a takeaway from this session, I would like to appeal to the security mindset; some kind of zero trust awareness for AI helps prevent a lot of problems.