4 ms·
Am I the only one who thinks that AI accountability is a fool's errand? The genie is out of the bottle and we know how to build these things now. I assume that
by 2OEH8eoCRo0 3y ago
Am I the only one who thinks that AI accountability is a fool's errand? The genie is out of the bottle and we know how to build these things now. I assume that AI will be put to every unethical/evil purpose possible (as humans often do) and it is only a matter of time.
Maybe it's still worth trying?
- alex_suzuki 3y agoI think it’s naive to assume otherwise. Relevant cartoon by The Economist: https://www.economist.com/the-world-this-week/2023/11/02/kals-cartoon https://www.economist.com/the-world-this-week/2023/11/02/kal...
- SiempreViernes 3y agoI don't understand what you mean by "accountability", I understand the word to mean the concept of actors paying a price if they don't take reasonable actions to protect others from potential future harm and compensate those that they have already harmed. Why would Microsoft creating an AI to help local authorities identify gay people in Chechnya and releasing it as open source mean Microsoft does not need to be held accountable?
- 2OEH8eoCRo0 3y agoI don't think that Microsoft would create an AI that does that. That's what I mean when I say we (the world) know how to build these AIs now. How would you prevent a world of people from creating unethical AIs? It's only a matter of time.
- SiempreViernes 3y agoThe same way we prevent people from doing bad stuff? Strong norms, clear laws, and active policing... it's not rocket science man.
- 2OEH8eoCRo0 3y agoHow do you prevent people from writing code or buying GPUs not just in the US but worldwide? You are missing my point entirely. The barrier to entry is so low (and getting lower) and policing difficulty so high it seems inevitable. We also passed laws against and police computer crimes and those are on the rise for the same reason- low barrier to entry and difficulty policing.