On Tuesday’s broadcast of CNN’s “Situation Room,” Sen. Mark Warner (D-VA) said that we need to beat China on AI and there should be international validation, but “we’ve not seen them try to race ahead on what’s called these frontier models.”
Warner stated that “one of the critiques people make, well, if we do any of this, China’s going to race ahead. Well, let me just point out — and I want America to win the AI race, I agree with that — but China’s already, for example, put in regulations for its own people that you can’t create a chatbot that will create a friendship with a young person. Don’t our kids deserve the same protection? China’s already put in place rules about AI using deepfakes and being able to fool people. Shouldn’t Americans have that?”
He continued, “And not to go totally nerdy, but China has been advocating more of what’s called open weights and open-source models. And they’ve been pretty darn effective at that. And how they did that is through a process called distillation, where they basically take our leading lab models and steal a lot of the information and then put out a product that’s 10% of the cost and 90% less energy usage. That’s a pretty compelling model. And we’ve not seen them try to race ahead on what’s called these frontier models. It would be better with an international validation. But the idea that China would not participate in that, I think, is not an — we, at least, ought to try.”
Follow Ian Hanchett on Twitter @IanHanchett
Read the full article here