> a convenient distraction from these unquestionably real issues we will face, maybe even a deliberate part of helping it happen?
Oh certainly. Not for any higher purpose, but remember that OpenAI lost its shit and threatened to leave the moment the EU wanted to look into actually regulating the risks of their product. The driver is money, and little else.
AI people like Sam Altman want a very specific kind of "AI regulation" - a mostly toothless one meant to address the hypothetical of a Skynet (which is honestly still pretty far out there) rather than the more realistic problems that come from their product.
Things like IP law (these LLMs rely on scraping lots of copyrighted data), profiling people (in an age where privacy is important to a lot of people), medical malpractice (I remember someone saying they fed a patient dossier into ChatGPT on this site... Dear Lord I hope that person got shafted by HIPAA) are all genuine problems/risks with LLMs that don't come from the models themselves but from the people using them. Thats the real danger and OpenAI wants absolutely no regulation in that area because it could hurt their bottom line.
Hence why they also almost all tend to follow the beliefs of Elezier Yudkowsky, whose entire life is basically to claim Rokos Basilisk is totally happening and the only option is to give him (and his friends) lots of money to ensure it so you won't be tortured by the AI for not bringing it about.
Crowing about evil AIs is a useful distraction because it plays well populistically (Skynet, Matrix, HAL 9000 are all good examples of "AI out of control" in the public consciousness) and it gives them easy backing from people who don't understand how these things work. The EU wasn't impressed and the AI act they passed addresses some actually meaningful concerns. It remains to see if US Congress/Senate will address it at all or if their business interests reign supreme.
Oh certainly. Not for any higher purpose, but remember that OpenAI lost its shit and threatened to leave the moment the EU wanted to look into actually regulating the risks of their product. The driver is money, and little else.
AI people like Sam Altman want a very specific kind of "AI regulation" - a mostly toothless one meant to address the hypothetical of a Skynet (which is honestly still pretty far out there) rather than the more realistic problems that come from their product.
Things like IP law (these LLMs rely on scraping lots of copyrighted data), profiling people (in an age where privacy is important to a lot of people), medical malpractice (I remember someone saying they fed a patient dossier into ChatGPT on this site... Dear Lord I hope that person got shafted by HIPAA) are all genuine problems/risks with LLMs that don't come from the models themselves but from the people using them. Thats the real danger and OpenAI wants absolutely no regulation in that area because it could hurt their bottom line.
Hence why they also almost all tend to follow the beliefs of Elezier Yudkowsky, whose entire life is basically to claim Rokos Basilisk is totally happening and the only option is to give him (and his friends) lots of money to ensure it so you won't be tortured by the AI for not bringing it about.
Crowing about evil AIs is a useful distraction because it plays well populistically (Skynet, Matrix, HAL 9000 are all good examples of "AI out of control" in the public consciousness) and it gives them easy backing from people who don't understand how these things work. The EU wasn't impressed and the AI act they passed addresses some actually meaningful concerns. It remains to see if US Congress/Senate will address it at all or if their business interests reign supreme.