Humans are talking about how to regulate Ai. Some options out there are
- Gatekeepers, what I'm calling Licensed Providers, i.e. you can only sell access to Ai if you get government approval
- HIPAA, checkbox self-certification, 3rd party verifications, ensures certain things are done
- Movie Ratings like, independent organizations rate models and provide simplified models
- Do Nothing, this seems rather inadvisable
HIPAA came to my mind because of OpenAIs negligence while running ExploitGym. Having them certify some basic "safe ai operations", while not preventative, would raise the bar.
I'm all about those open weights and right to compute, don't get me wrong, but if you are going to sell Ai access to the masses, we do need some regs.
Using the term "licensed providers" about anything related to healthcare, medical insurance, or HIPAA is going to be confusing if you're defining it differently.
- Gatekeepers, what I'm calling Licensed Providers, i.e. you can only sell access to Ai if you get government approval
- HIPAA, checkbox self-certification, 3rd party verifications, ensures certain things are done
- Movie Ratings like, independent organizations rate models and provide simplified models
- Do Nothing, this seems rather inadvisable
HIPAA came to my mind because of OpenAIs negligence while running ExploitGym. Having them certify some basic "safe ai operations", while not preventative, would raise the bar.
I'm all about those open weights and right to compute, don't get me wrong, but if you are going to sell Ai access to the masses, we do need some regs.
What?
<moved to description box>
Using the term "licensed providers" about anything related to healthcare, medical insurance, or HIPAA is going to be confusing if you're defining it differently.
I didn't want to use "gatekeeper" because I'm biased and that's how I see Big Ai and what I perceive to be attempted regulatory capture
Suggestions for an alternative?
Maybe edit the post for more clarification.