UK Launches Permanent AI Safety Institute to Lead in AI Policy

Lion Herald Newsletter | Issue #11 - Thursday November 2, 2023 | Lincoln, UK

November 02, 2023

Dear Lion Herald reader,

The United Kingdom is taking a significant step towards establishing itself as a leader in AI policy with the launch of the AI Safety Institute, a permanent global hub dedicated to testing the safety of emerging AI technologies. The formal announcement was made today by UK Prime Minister Rishi Sunak during the AI Safety Summit, marking a pivotal moment in the country's AI strategy.

The AI Safety Institute, which was informally unveiled in the lead-up to the summit, is set to be a crucial institution for evaluating the safety of new AI models. Ian Hogarth, an experienced investor, founder, and engineer who previously chaired the Frontier AI taskforce, will lead the institute. Furthermore, the renowned AI expert Yoshua Bengio will take charge of producing its first report.

While the government has not disclosed the exact funding allocation for the AI Safety Institute, it is anticipated that both government funding and contributions from industry players may be involved. The institute, operating under the Department of Science, Innovation, and Technology, is described as being "backed by leading AI companies," though the extent of this support remains to be clarified.

This significant development is accompanied by the announcement of the Bletchley Declaration that we covered yesterday, a new agreement endorsed by all the countries participating in the summit. The declaration commits nations to collective efforts in testing and assessing the risks associated with "frontier AI" technologies, including large language models.

Prime Minister Rishi Sunak in his speech emphasized the need for collaboration in AI safety testing, stating, "Until now, the only people testing the safety of new AI models have been the very companies developing them." He highlighted the importance of working together to ensure that AI models are tested for safety before their release.

While the UK's approach to regulating AI technology has been cautious, with a reluctance to impose regulations too early, some critics argue that more concrete actions are needed to ensure the safety and responsible development of AI.

Sunak defended the UK's approach, asserting that technology is advancing rapidly, and governments must ensure they can keep pace with developments. He stated, "Before you start mandating things and legislating for things... you need to know exactly what you're legislating for."

The AI Safety Summit, while emphasizing transparency in long-term AI efforts, held a series of closed-door plenary sessions involving key figures from government and industry. Notable attendees included CEOs of prominent AI companies and high-ranking government officials, such as UK Prime Minister Rishi Sunak and US Vice President Kamala Harris.

China, which had generated significant anticipation during the summit's first day, was notably absent from the closed plenaries on day two. Additionally, Elon Musk, the owner of X.ai (formerly known as Twitter), was not present during today's sessions, although he is scheduled for a fireside chat with Prime Minister Sunak on Musk's social platform, which is not expected to be a live broadcast.

As AI policy and safety discussions continue, the UK's establishment of the AI Safety Institute represents a significant step towards global leadership in responsible AI development and governance.

Thanks for reading,

Fabrice Iranzi, Lion Herald

Tech Headlines Today

UK-based vacation rental risk management startup Superhog has raised £5.5 million in Series A funding, led by investors including 6 Degrees Capital, Hambro Perks, and Solano Partners. This capital infusion will enable Superhog to expand its product and engineering teams, hire senior operations professionals, and establish international hubs in the United States and Australia. The company aims to enhance its proprietary risk assessment and protection technology while extending its global presence. Superhog's mission is to help property managers minimize risks in the short-term rental sector, and its solutions have reduced incident rates and driven property manager growth. The company boasts over 800 property management clients worldwide and annual recurring revenue exceeding £5 million.

The UK government has tripled its funding for supercomputers to support advanced AI research, allocating £300 million instead of the initially announced £100 million. Cambridge, chosen as the site for the second supercomputer, alongside Bristol, will host cutting-edge AI infrastructure. The supercomputers, "Dawn" in Cambridge and "Isambard-AI" in Bristol, will significantly bolster the UK's capabilities in AI research and technology. Dawn, powered by 1,000 water-cooled chips, will focus on climate modeling, healthcare, and fusion energy, while Isambard-AI, with over 5,000 Nvidia chips, aims to drive AI and scientific breakthroughs. The move reflects the UK's commitment to lead in AI technology safely and facilitate groundbreaking research. However, concerns have been raised about ensuring adequate electricity supply for the supercomputers' power needs.

Thanks for joining us in this edition of the Lion Herald Newsletter. We hope you've found inspiration and valuable insights to fuel your tech and innovation journey. Remember, every idea counts, and together, we can turn them into real-world solutions. Stay curious, stay innovative, and stay tuned for more exciting stories in our next issue!

Contact | Lion Herald Ltd. Tel : +447706122028, Office: 13-14 Silver Street, Lincoln LN21DY (Mosaic Digital Hub -Thomas Parker House)