You’re modeling government as a single coherent actor—and I think that’s the most critical mistake. That’s not to say they are incompetent, just that governments aren’t actually looking at what companies do to decide how to respond. (And many would say this is a feature, not a bug!)
Sorry if my post made it seem that way, but I don’t feel like I’ve been thinking of it that way. In fact, it’s sort of worse if it’s not a single actor; many different departments could have done something about this, but none of them seemed to take public action.
I’m not sure how to understand your second sentence exactly. It seems pretty different from your first sentence, from what I can tell?
A multi-actor system is constrained in ways that a group of single actors are not. Individual agencies can’t do their own thing publicly, and you can’t see what they are doing privately.
For the agencies that do pay attention, they can’t publicly respond—and the lack of public monitoring and response by government agencies which can slap new regulations on individual companies or individuals is what separates a liberal state from a dictatorship. If US DOD notices something, they really, really aren’t allowed to respond publicly, especially in ways that would be seen as trying to interfere with business or domestic policy. If NSA or the FBI notices something, they can only enforce extant laws, and are limited in their legal ability. And agencies which can respond, like the FTC, are in fact already working on drafting regulations for relevant applications of AI. (And yes, Congress could act to respond, but it’s really fundamentally broken.)
You’re modeling government as a single coherent actor—and I think that’s the most critical mistake. That’s not to say they are incompetent, just that governments aren’t actually looking at what companies do to decide how to respond. (And many would say this is a feature, not a bug!)
Sorry if my post made it seem that way, but I don’t feel like I’ve been thinking of it that way. In fact, it’s sort of worse if it’s not a single actor; many different departments could have done something about this, but none of them seemed to take public action.
I’m not sure how to understand your second sentence exactly. It seems pretty different from your first sentence, from what I can tell?
A multi-actor system is constrained in ways that a group of single actors are not. Individual agencies can’t do their own thing publicly, and you can’t see what they are doing privately.
For the agencies that do pay attention, they can’t publicly respond—and the lack of public monitoring and response by government agencies which can slap new regulations on individual companies or individuals is what separates a liberal state from a dictatorship. If US DOD notices something, they really, really aren’t allowed to respond publicly, especially in ways that would be seen as trying to interfere with business or domestic policy. If NSA or the FBI notices something, they can only enforce extant laws, and are limited in their legal ability. And agencies which can respond, like the FTC, are in fact already working on drafting regulations for relevant applications of AI. (And yes, Congress could act to respond, but it’s really fundamentally broken.)
An, that’s really good to know… and kind of depressing. Thanks so much.