Rishi Sunak is preparing to launch AI Summit and calls for UK-backed bots
The chief executive of an important government artificial intelligence (AI), supplier, has called on Britain to build a competitor to ChatGPT.
Marc Warner, the CEO of Faculty, which has been contracted by government officials to do research in advance of the AI Safety Summit said that such a system could be “a crucial capability for national defence and security”.
Government advisors are debating if Britain should develop , a sovereign capability, to rival OpenAI’s ChatGPT instead of relying on US tech giants who are quickly dominating the industry.
Chatbots, such as ChatGPT (built on a technology called a large-language model), have sparked a public interest in AI. They can analyze text quickly and provide natural language answers to questions. As technology advances, we are seeing ever more powerful capabilities, like generating code.
These tools can be used to automate certain tasks, and there are fears that millions will lose their jobs.
Warner, who advised the Government on the Covid pandemic said that the UK would like to have “as many controls as possible” in the defence context over the technology.
A report on ChatGPT, co-authored by GCHQ, stated that the technology had “promising prospects” for intelligence work. However, it could only replace an “extremely Junior Analyst”.
US technology company Palantir has shown how chatbots can be used to provide real-time analyses and recommendations for commanders on the battlefield.
He said that he thought it was best to focus on national security and defense as the most effective way of achieving the highest level of sovereign capability.
Warner, 39, is an entrepreneur who co-founded Faculty 10 years ago. The start-up worked closely with both the NHS and the Government on the implementation of AI technologies.
Rishi Sunak’s remarks coincide with the UK AI Safety Summit, which Rishi will be hosting at Bletchley Park.
All three are expected to attend: Sam Altman, chief executive officer of OpenAI; Dario Amodei, of the rival Anthropic; and Demis Hassabis, of Google DeepMind.
Elon Musk, Google, Amazon, and Meta’s artificial-intelligence company xAI, as well as executives from Google, Amazon, and Google will be present at the event.
The Government’s “Frontier AI” has been the focus of the preparations for the summit.
Sunak, who spoke on Thursday about the summit, said that it would also address concerns over disinformation and deep-fakes, as well as the threat of , a malicious “superintelligence”, which mankind could “lose complete control of”.
Experts are fervently divided over such claims . On Tuesday, 23 researchers, including a UK Government adviser, published an article warning that more powerful AI systems may trigger the “extinction” of humanity.
Some have dismissed these warnings as unimportant, a distraction, and far beyond the capabilities that humans will likely develop in the future.
Warner stated that there is a need to maintain balance between the long-term threat and the opportunities presented by technology. He said, “We want nuance.”
Warner says that despite this, AI’s potential risks are “very real”.
He said, “I’m in the camp that believes we should be concerned about long-term risk.” “I don’t want another global warming situation where we start to think about it only when it gets really bad.”
Faculty was hired last month by the Department for Science, Innovation and Technology (DSTI) to better understand potential threats.
The work included stress testing AI systems through a red teaming process to determine how they could be used by lone-wolf terrorists, and if they could amplify those who wish to cause mass devastation.
Warner, a long-time advisor to the Government, was called to assist No10 when the Covid pandemic began, along with his younger brother Ben.
They worked closely with Dominic Cummings at the time, Boris Johnson’s chief advisor, to help Britain’s first response to the crisis. They gathered and analysed key health data.
Cummings said that if he had been the Prime Minister in 2021 he would have “put Marc Warner… in charge” of this entire thing.
It is not surprising that he has backed him, as Faculty and Cummings worked together on the Brexit Vote Leave 2016 campaign. The company does not work on political campaigns anymore. Warner, who was a prominent figure in the Government’s AI Council until it was disbanded this summer, remains close with those working now as part of UK’s Frontier AI Taskforce.
Warner stated that as AI systems become more powerful there will likely be a requirement for licensing or vetting those who access the technology.
Warner said: “It is unlikely that society would tolerate long-term people working on dangerous things without sensible safeguards.”
OpenAI is raising money at a valuation of $90bn to keep out competitors.
Warner said that there are still “huge” benefits from the large language model being developed by OpenAI. AI can be used, for example, to enhance the work of doctors by filtering through decades worth of medical literature and notes to produce personalised diagnoses.
He said, “That’s going to be brilliant.” “And it’s about doable right now.”
Faculty has already helped the NHS to use ChatGPT for creating synthetic patient data that can be used in simulations.
In order to compete with the US, the UK must ensure that it has access to – the powerful GPUs used in AI technologies such as ChatGPT. The government is planning to purchase around 5,000 processors.
It may not be sufficient. Warner says that commercial organisations are considering 30,000 GPUs for now and 100,000 for the future. “Those numbers are out of control”.
Warner stated that despite the challenges, winning the AI race would be critical to reverse the UK’s stagnant GDP. Abdicating leadership would mean creating a UK generation “that is poorer than their parents”.
Another debate is looming over the AI Summit next week – should China even be invited? Liz Truss, the former British prime minister, wrote Rishi Sunak on Thursday to urge him to withdraw Beijing’s invitation.
Warner, on the other hand, said that the global nature AI risks warranted bringing China to a top-table.
He said: “It would be a bad idea to exclude China from international discussions.”
These questions are about the humanity as a collective.