I'm on the same side as you, vehemently fuck generative AI, however the difference between some LLM output incorrectly telling you that you have cancer and a shitty coder vibe coding with severe vulnerability is the ability to place blame, there's a person who's job it was to make sure info is secure and the vibe code actually works. There is no singular person who is traceably culpable for shitty AI output telling you bad info.