Lots of speculation about AI becoming a sentient skynet and killing off humanity (I still think consciousness is a barrier to real artificial intelligence).
My worries about AI are more mundane, but no less threatening. AI as of now is essentially an advanced search engine. But instead of providing pointers to answers, as a traditional search engine does, it provides the answer itself.
If AI is skewed by input bias or editorial control, the answer it provides will be a semi-truth or an outright lie. As people become more reliant on AI, the biases and lies will lead to society-wide competency crises.
If AI is an accurate assessment of reality, it is likely to lead to mass unemployment when expertise becomes accessible to the layman. White collar intermediary jobs would take the biggest hit.
If AI imagery becomes indistinguishable from reality (we are almost there), the masses can be emotionally manipulated to support policies or politicians that go against their interests.
AI will empower hackers and render security and privacy controls obsolete. When that happens, people will flee the internet just to have a sense of control over their personal lives. No data will be safe from prying eyes.
Ultimately I see AI accelerating the pace toward Zero Social Trust in societies that are already experiencing a drastic decline in social trust from racial diversity. I also see it turbocharging the surveillance state, which will, when fully deployed and operational, mentally enslave the masses in chains of fear and paranoia.
