OpenAI, DeepMind Will Open Up Models To UK Government (politico.eu) 17
Google DeepMind, OpenAI and Anthropic have agreed to open up their AI models to the U.K. government for research and safety purposes, Prime Minister Rishi Sunak announced at London Tech Week on Monday. From a report: The priority access will be granted in order "to help build better evaluations and help us better understand the opportunities and risks of these systems," Sunak said. The announcement came in a speech that championed the promise of AI to transform areas such as education and healthcare and heralded the U.K.'s potential as an "island of innovation."
"AI is surely one of the greatest opportunities before us," said Sunak. By combining AI models with the power of quantum, "the possibilities are extraordinary," he marvelled. "But we must and we will do it safely," he continued. "I know people are concerned."
"AI is surely one of the greatest opportunities before us," said Sunak. By combining AI models with the power of quantum, "the possibilities are extraordinary," he marvelled. "But we must and we will do it safely," he continued. "I know people are concerned."
Re: Look-good gesture, ultimately pointless (Score:3)
They don't undersrand how email works. This is an empty gesture.
Re: (Score:1)
it's probably exactly what they're banking on. They won't actually open up their systems either and they're probably correctly guessing that nobody will be able to tell, at least not the kind of people who end up getting these kinds of jobs in the government. And if they can tell, they probably have been incentivized not to say.
Re: (Score:2)
They probably will give them a snapshot of the models. By the time anyone competent sees it (probably one of their competitors) it will be obsolete.
Training data (Score:5, Insightful)
Re: (Score:3)
Re: (Score:3)
I'm not sure what they are looking for. I'd be surprised if the code for GPT4 is any more than maybe a thousand lines of code and a bunch of off the shelf libraries.
Seriously, transformers are remarkably simple algorithms for what they actually do. Theres probably more code in the ChatGPT front end.
But understanding what is happing in those giant inscrutable matrixes, well thats another question. The algorithms are simple, the data on the other hand, not so much, and clues are starting to come in as to what
Government? (Score:1)
They may as well be explaining a magic trick to a dog.
Re: (Score:3)
Re:Regulatory controls will only backfire. (Score:4, Insightful)
I'd expect Japan to be more accepting than China. Japan is desperate for better robots, and needs to cope with an aging workforce, and doesn't want immigrants.
"safety purposes" (Score:1)
I needed a good laugh. Did they have a real choice? This is a warning.
The last thing governments want is for AI to tell the truth. The number of lies told by government and business over the past 5 years is staggering.
The government is fearful that AI can squash a lie or narrative before it gets traction.