Prime Minister Rishi Sunak is preparing to host the AI Safety Summit at Bletchley Park in November with a strong focus on the UK's evolving AI policy and its global implications. The urgent question: can the UK leverage this summit to establish itself as a global leader in ethical AI governance while still debating its own AI policy internally?
The recently published interim report on AI governance from the UK's Science and Technology Select Committee, led by Greg Clark MP, stressed the urgency for the UK to take leadership in AI regulation. The report said, "The AI white paper should be welcomed as an initial effort to engage with this complex task, but its proposed approach is already risking falling behind the pace of development of AI. This threat is made more acute by the efforts of other jurisdictions, principally the European Union and United States, to set international standards."
The report warns that without prompt action to establish regulatory frameworks, the country risks falling behind in the global AI arena. The report is critical of the government's current "pro-innovation" policy, which relies on existing regulators without giving them additional resources or powers. Mr Clark warned that the upcoming King's Speech may be the last chance for legislative measures before the next General Election.
While the report acknowledges the initial efforts of the Government's White Paper on AI, it warns that its suggested approach might lack the required legislative impact, that is despite the White Paper referring to a "statutory duty" for regulators.
The report notes that governments, AI firms, and researchers must collectively address the twelve challenges AI poses:
- Bias Challenge: Addressing how AI can introduce or perpetuate societal biases.
- Privacy Challenge: Managing AI's ability to identify individuals and use their personal information in undesired ways.
- Misrepresentation Challenge: Tackling AI-generated material falsely portraying someone's actions, beliefs, or character.
- Access to Data Challenge: Overcoming the concentration of large datasets in the hands of few organisations.
- Access to Compute Challenge: Ensuring that a few do not monopolise the high computational power required for AI development.
- Black Box Challenge: Solving the lack of transparency in how some AI models make decisions.
- Open-Source Challenge: Balancing the pros and cons of making AI code openly available versus proprietary.
- Intellectual Property and Copyright Challenge: Defining and enforcing the rights of original content creators when AI uses their work.
- Liability Challenge: Determining who is responsible for harm caused by third-party use of AI tools.
- Employment Challenge: Preparing for job disruption due to increasing AI automation.
- International Coordination Challenge: Collaborating globally to develop AI governance frameworks.
- Existential Challenge: Addressing concerns that AI could significantly threaten human existence and national security.
The Committee's report sets the stage for the upcoming AI Safety Summit as the UK grapples with shaping global AI regulations and standards. The historic venue of Bletchley Park for the AI Safety Summit is a nod to the UK's pioneering role in computing and AI. It was here that Alan Turing, considered the father of modern computing, cracked the German Enigma code during WWII. Turing also proposed the Turing Test, a cornerstone in AI that gauges a machine's ability to display human-like intelligence. Hosting the summit is not just a nod to the UK’s computing history, but also a critical reminder that ethical considerations in technology have been crucial since the field's inception.
The Prime Minister aims to make the UK the hub for AI governance, aligning with OpenAI's CEO, Sam Altman, who has proposed for a global AI regulatory body similar to the International Atomic Energy Agency (IAEA), with the Prime Minister wanting it headquartered in the UK.
Yet, as the UK aims to steer the global conversation, it faces a complex geopolitical landscape concerning China's role. While including China could lend legitimacy to a worldwide AI body, it also presents security risks and potential strains on relations with allies like Japan and the European Union. The US seems more amenable to China's involvement, complicating the UK's diplomatic efforts. This has been compilated further by a UK Parliament researcher being arrested under anti-espionage laws for allegedly spying for China. He has denied the allegations. The arrests have nonetheless reignited debates among UK politicians about the country's stance towards China, which could complicate the upcoming summit and future regulation.
Navigating the complexities of global AI governance, the UK could consider a "two-track" approach to maintain both inclusivity and security. The first track would involve a broad international forum that includes all nations, offering legitimacy but potentially introducing security risks and straining alliances with democratic countries. In parallel, a second forum exclusively for democratic nations would focus on shared values like transparency and individual rights, facilitating quicker consensus-building. This dual-forum strategy mirrors the International Atomic Energy Agency's successful Cold War-era model, allowing for broad cooperation while upholding democratic principles. Such a compromise aims to harmonise divergent national interests, providing a framework for mutually beneficial international AI governance. The UK can be a leader in this, but its role is far from assured.