Statement on AI Risk

I find Robert Miles worryingly plausible when he says (about 12:40 into the video) "if you have a sufficiently powerful agent and you manage to come up with a really good objective function, which covers the top 20 things that humans value, the 21st thing that humans value is probably gone forever"

Just take the recent events alone - COVID-19 would not have been as much of a threat to humanity if some people wouldn't have built echo chambers on the internet with tremendous influence over others where they would share their unfounded conspiracy theories and miracle cures (or miracle alternatives to protecting oneself).

But there is a lot more. The data collection through the internet has enabled politicians who have no clue how to lead to be elected through just saying the right things to the largest demographic they can appeal to. Total populism and appeasing the masses has always been an effective strategy for politicians, but at least they could not execute it effectively. Now, everyone with enough money can. And this definitely stifles human progress and enshrines a level of regression in our institutions. Potentially dangerous regression, especially when it involves prejudice against a group or stripping away rights, just because people talk about it in their DMs on social media and get binned into related affinity buckets for ads.

Then there is the aspect of the internet creating tremendous time-wasters for a very large proportion of the population, robbing humanity of at least a million man-years of productivity a day. It is too addictive.

It has also been used to facilitate genocides, severe prejudice in large populations, and other things that are extremely dangerous.

High risk? Maybe not. A risk, though, for sure. Life was significantly more positive, happier and more productive before the internet. But the negative impact internet has had on our lives and human progress isn't all that it could have had. When a senile meme president gets the nuclear codes thanks in part to a funny frog picture on the internet, I think that is enough to say it poses a risk to extinction.

The right-wing is tired of the California ideology, is invested in the primary and secondary sectors of the economy, and has learned to mistrust claims that the technology industry makes about itself (regardless of those claims are prognostications of gloom vs bloom).

The left-wing thinks that technology is the driving factor of history, is invested in the tertiary and quaternary sectors of the economy, and trusts claims that the technology industry makes about itself. Anytime I see a litany of "in 10 years this is gonna be really important" I really just hear "right now, me and my job are really important".

The discussion has nothing to do with whether AI will or will not change society. I don't think anyone actually cares about this. The whole debate is really about who/what rules the world. The more powerful/risky AI is, the easier it is to imagine that "nerds shall rule the world".