Yeah, Rishi, it's AI that'll make Britain great again
DeepMind, OpenAI, Anthropic promise to give Blighty priority access to their models for safety research
UK prime minister Rishi Sunak believes taking a leading role in AI safety research is key to boosting Britain's innovation capability and growing its economy.
On Monday, in a speech kicking off the tenth London Tech Week, Sunak said the country must act quickly to retain its position as one of the world's top tech capitals, and work towards making it the best place to launch and invest in startups.
"If our goal is to make this country the best place in the world for tech, AI is surely one of the greatest opportunities before us," he opined. "The possibilities are extraordinary. But we must – and we will – do it safely.
"I know people are concerned. The very pioneers of AI are warning us about the ways these technologies could undermine our values and freedoms through to the most extreme risks of all. And that's why leading on AI also means leading on AI safety."
Sunak believes the technology needs to be developed more safely for it to be adopted on a wide scale.
The announcement comes at a time when experts have signed public letters claiming that AI poses a risk to humanity much like nuclear war or pandemics, and that development in training the largest and most capable models should be paused. Lawmakers around the world are also stepping up efforts to regulate AI and tackle issues of bias and misinformation – well, some of them at least.
Anything to deflect from the wholesale scraping of the internet – from books to social media posts to articles to photos to music – to train these models on everyone else's work, right?
- Europe to vote on AI laws with potential 7% revenue fines
- Betting on AI leads to more educated juniors, fewer mid-level bosses – study
- AI weapons need a safe back door for human control
- Why can't Nvidia boss Jensen Huang escape the Uncanny Valley that makes AI feel icky?
Sunak said the £100 million ($125 million) investment into the Foundation Model Taskforce will be spent on AI safety. "We're dedicating more funding to AI safety than any other government," he claimed.
He also announced that Google DeepMind, OpenAI, and Anthropic had committed to giving "early or priority access" to their models for research into safety risks to "help build better evaluations" and "help better understand the opportunities and risks of these systems."
To attract more AI talent, Sunak reiterated that the British government has pledged £900 million ($1.12 billion) into computing resources – including a possible supercomputer to train the large language models. He also wants to introduce the High Potential Individual, Global Talent and Scale-up Worker visas to make it easier for people with AI expertise to live and work in the UK.
Last week, he announced plans to host the first global summit on AI safety later this year, and met with US president Joe Biden in the White House. They spoke about how the two countries can collaborate on AI in the future to strengthen trade and grow their economies. ®