I appreciate that. China is a threat in a lot of ways. I wouldn't be surprised if we wind up in conflict over Taiwan. Are they a threat in AI? Maybe-but Prof Miller and I both think the leadership prizes regime survival and that means total control over something as potentially powerful as AI. They're not going to rush into that with the reckless abandon we are if there is a tiny chance it might destabilized the CCP or Xi's grip on power. They've got the most restricted internet on the planet (other than North Korea). They're very cautions about outside influence and tech. I think "But China" is mostly a move to get support from the DOD etc for "acceleration" in AI development.
There are so many fascinating bits in the two parts. It is hard to pick out any one thing, but the discussion of China was perhaps the most unexpected. I wish everyone would read that when journalists and politicians go on and on about the threat of China. There is a threat, but it is probably not what most people think it is.
"Perhaps it should be stigmatized like other really bad things get stigmatized, like pedophilia, or human trafficking, or incredibly dangerous synthetic drugs, right? "
Might as well stigmatize, but getting the press (and even prosecutors) to cooperate will be a big problem. And despite stigmatization and regulation, illegal drugs are rampant, as is human trafficking, and pedophiles can obtain outlawed images and videos without too much effort. Lawsuits by people impersonated will have good grounds to prevail, but new fakes will pop up from other vendors, potentially too many to target with libel suits and takedowns. It's depressing. Money always wins.
I'd say, "money usually wins." But if you can build in a culture of AI skepticism, stigmatization does have a chance. The "shogun" period of feudal Japan was extended by decades after the Shogunate decided, on balance, they didn't like guns, even though they offered military advantage. There are plenty of societies that looked at a particular innovation and thought, "You know what? Thanks, but we don't want it." It is definitely an uphill slog to slow down AI and try to force AI companies to make safety a priority. But it is doable, once people viscerally understand the extent of the potential threat. Right now, its still science fiction to most folks.
I appreciate that. China is a threat in a lot of ways. I wouldn't be surprised if we wind up in conflict over Taiwan. Are they a threat in AI? Maybe-but Prof Miller and I both think the leadership prizes regime survival and that means total control over something as potentially powerful as AI. They're not going to rush into that with the reckless abandon we are if there is a tiny chance it might destabilized the CCP or Xi's grip on power. They've got the most restricted internet on the planet (other than North Korea). They're very cautions about outside influence and tech. I think "But China" is mostly a move to get support from the DOD etc for "acceleration" in AI development.
There are so many fascinating bits in the two parts. It is hard to pick out any one thing, but the discussion of China was perhaps the most unexpected. I wish everyone would read that when journalists and politicians go on and on about the threat of China. There is a threat, but it is probably not what most people think it is.
"Perhaps it should be stigmatized like other really bad things get stigmatized, like pedophilia, or human trafficking, or incredibly dangerous synthetic drugs, right? "
Might as well stigmatize, but getting the press (and even prosecutors) to cooperate will be a big problem. And despite stigmatization and regulation, illegal drugs are rampant, as is human trafficking, and pedophiles can obtain outlawed images and videos without too much effort. Lawsuits by people impersonated will have good grounds to prevail, but new fakes will pop up from other vendors, potentially too many to target with libel suits and takedowns. It's depressing. Money always wins.
I'd say, "money usually wins." But if you can build in a culture of AI skepticism, stigmatization does have a chance. The "shogun" period of feudal Japan was extended by decades after the Shogunate decided, on balance, they didn't like guns, even though they offered military advantage. There are plenty of societies that looked at a particular innovation and thought, "You know what? Thanks, but we don't want it." It is definitely an uphill slog to slow down AI and try to force AI companies to make safety a priority. But it is doable, once people viscerally understand the extent of the potential threat. Right now, its still science fiction to most folks.