[IngvarLynn]

>AI hacked nothing. We've been doing it to ourselves for ages. This is just another way of doing it

except much more effective.

Humanity barely managed to get over (if so) previous mind-viruses like bible and communism. Partially because of massive discussion around these topics. With current LLMs (not future superhuman AGI) it's possible to spoon-fed a personally tailored mind-virus to each internet user. And there would be no discussion just because every instance is so original. (In practice it's often enough to just rename all the keywords, as proven by numerous psycho/business-cource gurus.)

Previously: https://news.ycombinator.com/item?id=24366308

[alganet]

If the story is too personal, you can't share it and influence others. In your analogy, is a virus that can't spread.

A pop song can feel deeply personal to many people, precisely because it's vague and not specific. The composer doesn't need to know the listener at the individual level. It needs to know an audience.

The combination of grooming a specific audience and then targetting it broadly is much more likely to work than individual propaganda pieces, and much more effective because the affected are inclined to influence others within the audience. It's already being done in a massive scale.

Yes, there are concerns about the targetting individuals with AI. Whistleblowers, activists and so on are in serious risk. These are likely to have some antibodies though, and that is not what the article is talking about.

The best defence in my opinion is better education, skepticism and proper critical thinking. Exactly the same stuff that prevents people from being targeted by the old stuff.