@prefec2 I think, we have conflicting agendas too. AIs owned by right-wing tech billionaires or military or dictators will have other goals, then the ordinary people. We had AIs that were xenophob or tried to eleminate the user to reach a goal.
Every human lives with some cognitive dissonance, but I don't know, if computers can do that too.
@GambaJo true. On the plus side HAL 9000 had a personality and was self aware. Current LLMs are not self-aware. On the negative side, they are statistical systems that reproduce what was inputted into them. And there applies the rule: Bullshit in, bullshit.out. And this also means racism in, racism out.
So yes they are dangerous, but for other reasons.
In general XKCD had this pretty well covered in the following cartoon (called Robot Uprising) https://xkcd.com/1968/
@GambaJo For me the key essence of the cartoon is:
A lot of people worry of the wrong fictional problem, while the real issue is nearer, as we currently work on that, and darker.
Regarding climate change, a lot think we are just fine, thinking of "we have to save the planet" as nice to have. However, the planet will be fine, but we will no longer be compatible with the living conditions.
Add comment