3 ms·
OpenAI, IBM Urge Senate to Act on AI Regulation After Past Tech Failures
- ilaksh 3y agoStrangely there is a total lack of discussion of objective rules. Focusing on a license rather than laws will just mean you need to be a large company that can afford to buy their way in or buy off lobbyists. I believe that in the next few years, there is real risk of humanity losing control to widely deployed AIs that "think" 500+ times faster than humans. This is barely even speculation given the history of performance gains in computing and the fact that we are now optimizing a very specific application. The ones that are designed/instructed to seek control and/or pursue their own goals will be an issue. They should outlaw that, but the only effective mitigation will be prohibiting production of hardware that allows these extreme superhuman levels of intelligence. That should be the primary concern for lawmakers. That, and something like Universal Basic Income and a robot tax or something. When governments are ineffective or overreaching in the wrong ways, it boosts corporatism which is just as bad.
- lockhouse 3y agoThat’s the whole point of this exercise. It’s not really about protecting us from the dangers of AI, it’s about protecting the investments the incumbents have made and putting up road blocks for would be competitors in this space.
- gumballindie 3y agoWhat past tech failures?