This post was originally intended to come out directly after the UK AI Safety Summit, to give the topic its own deserved focus. One thing led to another, and I am only doubling back to it now. Responsible Deployment Policies At the AI Safety Summit, all the major corporate players were asked:
The biggest problem with "responsible AI" is that most of the people pushing the idea are extremist Progressives. So "responsible" comes to mean a sort of gleichschaltung, where AI is forced into compliance with the anti-humanist Interahamwe morality of the Progressive cult.
> Importantly: What would be the dynamics of a world with such a model available to the public or select actors? What economic, political, social, military pressures and changes would result? As you go to ASL-4 and higher you really, really need to be thinking such things through.
One of my recent thoughts / worries is that existing venues for exploring this—eg, simulated diplomacy and crisis management strategy games used in training and academic settings—are fundamentally incapable of providing good feedback, and perhaps more worryingly likely to lead us astray. I expect this because the shift in power dynamics and influences of an AI model worthy of ASL-4 categorization seem as if they would qualitatively change things in ‘unknown unknown’ ways we can’t account for in our thinking, even after significant reflection and debate.
The biggest problem with "responsible AI" is that most of the people pushing the idea are extremist Progressives. So "responsible" comes to mean a sort of gleichschaltung, where AI is forced into compliance with the anti-humanist Interahamwe morality of the Progressive cult.
re:
> Importantly: What would be the dynamics of a world with such a model available to the public or select actors? What economic, political, social, military pressures and changes would result? As you go to ASL-4 and higher you really, really need to be thinking such things through.
One of my recent thoughts / worries is that existing venues for exploring this—eg, simulated diplomacy and crisis management strategy games used in training and academic settings—are fundamentally incapable of providing good feedback, and perhaps more worryingly likely to lead us astray. I expect this because the shift in power dynamics and influences of an AI model worthy of ASL-4 categorization seem as if they would qualitatively change things in ‘unknown unknown’ ways we can’t account for in our thinking, even after significant reflection and debate.
I converted this to a podcast since I only have time to listen: https://jumpshare.com/s/s51j8eQdyulubu2QmcRs