Artificial intelligence proposed to be licensed as drugs or nuclear energy: “Two years left”

Labor said the UK should ban technology developers from working on advanced AI tools unless they have a license to do so.

Ministers should impose much stricter rules on companies training their AI products on large datasets like the one OpenAI uses to build the ChatGPT chatbot, Lucy Powell, Labor Party digital spokesman, told The Guardian.

Her comments, writes The Guardian, came amid a rethinking at the highest government levels of how to regulate the rapidly changing world of artificial intelligence, and British Prime Minister Rishi Sunak acknowledged that AI could pose an “existential” threat to humanity.

One of the government’s artificial intelligence advisers also said on Monday that humanity could only have two years before AI can outsmart humans, the latest in a series of dire warnings about the threat posed by rapidly advancing technology.

Lucy Powell says: “What really worries me is the lack of any regulation of large language models, which can then be applied to a whole range of artificial intelligence tools, no matter how they are created, how they are managed or how they are controlled.”

She proposed licensing artificial intelligence in a similar way to medicines or nuclear power, which are regulated by government agencies at arm’s length. “This is the model that we need to think about when you have to have a license to make these models,” she said. “I think these are good examples of how this can be done.”

Two months ago, the UK government published a white paper on artificial intelligence that detailed the possibilities that cutting-edge technology could provide, but said relatively little about how to regulate it.

Since then, a number of developments, including advances in ChatGPT and a series of dire warnings from industry insiders, have sparked a rethink at the very top of the British government, with ministers now rushing to update their approach. Sunak will travel to Washington this week, where he will argue that the UK should be at the forefront of an international effort to write a new set of guidelines to guide the industry.

Labor is also rushing to finalize its own advanced technology policy. Lucy Powell, who is speaking to industry insiders at the techUK conference in London on June 6, said she believes the destruction of the UK economy could be as drastic as the de-industrialization of the 1970s and 1980s.

Labor leader Keir Starmer is expected to give a speech on the subject during London Tech Week. Starmer will host a shadow cabinet meeting next week at one of Google’s UK offices, giving shadow ministers the opportunity to speak with some of the company’s top AI executives.

Lucy Powell said that instead of banning certain technologies, as the EU did with tools like facial recognition, she thinks the UK should focus on regulating how they are developed.

Products such as ChatGPT are built with learning algorithms from vast databases of digital information. But experts warn that if these datasets contain biased or discriminatory data, then the products themselves may be indicative of those biases. This could have a side effect, for example, on employment practices if artificial intelligence tools are used to make hiring and firing decisions.

Powell says, “Bias, discrimination, surveillance – this technology can have many unintended consequences.”

She argued that by forcing developers to be more open about the data they use, governments could help mitigate those risks. “This technology is advancing so fast that it requires active government intervention, not hands-off.”

Matt Clifford, chairman of the Advanced Research and Invention Agency, set up by the government last year, said on Monday that artificial intelligence is advancing much faster than most people think. He said that this could already be used to launch biological weapons or large-scale cyber attacks, adding that humans could quickly outpace the technology they created.

Speaking to TalkTV’s Tom Newton Dun, Clifford said: “It’s certainly true that if we try to create artificial intelligence that is smarter than humans and we don’t know how to control it, that creates the potential for all sorts of risks now and in the future. . So I think there are a lot of different scenarios to worry about, but I certainly think it’s right that this should be very high on the politicians’ agenda.”

When asked when that might happen, he added: “No one knows. There is a very wide range of predictions among AI experts. I think two years will be the most optimistic in this regard.”