The AI Safety Summit 2023 was an international conference focused on the safety and regulation of artificial intelligence. Organized by the British government, it was held in November 2023 at Bletchley Park, Milton Keynes, England. The event was the first ever global summit on artificial intelligence, bringing together government leaders, technology executives, and researchers to address the risks associated with advanced AI systems. The summit led to the release of the Bletchley Declaration, which focused on "identifying AI safety risks of shared concern" and "building respective risk-based policies" to "ensure that the benefits of the technology can be harnessed responsibly for good and for all."
The summit marked a significant moment in the global conversation about Artificial intelligence governance, as it was the first time that major world powers, including the United States and China, agreed on a joint statement regarding AI safety. The event also underscored the growing importance of Large language model and Generative AI technologies, which had advanced rapidly in the years leading up to the summit.
Background
The prime minister of the United Kingdom at the time, Rishi Sunak, made AI one of the priorities of his government, announcing that the UK would host a global AI Safety conference in autumn 2023. The announcement came amid increasing public and governmental concern about the potential risks of AI, including its use in disinformation, cyberattacks, and autonomous weapons. Sunak's government sought to position the UK as a leader in AI safety regulation, building on the country's historical contributions to computing and its strong academic and industrial AI sectors.
The summit was planned as a high-level diplomatic event, with invitations extended to heads of state, technology company CEOs, and leading AI researchers. The British government emphasized the need for international cooperation, arguing that AI risks were global in nature and required coordinated responses.
Venue
Bletchley Park was a World War II codebreaking facility established by the British government on the site of a Victorian manor and is in the British city of Milton Keynes. It has played an important role in the history of computing, with some of the first modern computers being built at the facility. The choice of venue was symbolic, as Bletchley Park is widely regarded as the birthplace of modern computing, where figures like Alan Turing worked on breaking the Enigma code.
The venue's historical significance was highlighted by speakers and attendees, who drew parallels between the codebreaking efforts of the 1940s and the contemporary challenge of understanding and controlling AI. The setting also served to remind participants of the UK's legacy in technological innovation, from the early days of computing to the development of Transformer (architecture) architectures that underpin modern AI systems.
Outcomes
28 countries at the summit, including the United States, China, Australia, and the European Union, issued an agreement known as the Bletchley Declaration, calling for international co-operation to manage the challenges and risks of artificial intelligence. The Bletchley Declaration affirms that AI should be designed, developed, deployed, and used in a manner that is safe, human-centric, trustworthy and responsible.
Emphasis has been placed on regulating "Frontier AI", a term for the latest and most powerful AI systems. Concerns that have been raised at the summit include the potential use of AI for terrorism, criminal activity, and warfare, as well as existential risk posed to humanity as a whole. The declaration did not set specific regulatory targets but established a framework for future discussions, including a commitment to hold further summits.
In addition to the declaration, several national initiatives were announced. The president of the United States, Joe Biden, signed an executive order requiring AI developers to share safety results with the US government. The US government also announced the creation of an American AI Safety Institute, as part of the National Institute of Standards and Technology. These actions were seen as a direct response to the summit's emphasis on transparency and accountability in AI development.
The tech entrepreneur Elon Musk and Sunak did a live interview on AI safety on 2 November on X, the social media platform owned by Musk. The interview, which was broadcast to a wide audience, covered topics such as the potential for AI to surpass human intelligence and the need for regulatory oversight. Musk, who had previously called for a pause on AI development, reiterated his concerns about existential risks.
Notable attendees
The following individuals attended the summit:
- Rishi Sunak, Prime Minister of the United Kingdom
- Kamala Harris, Vice President of the United States
- Charles III, King of the United Kingdom (attending virtually)
- Elon Musk, CEO of Tesla, owner of X, SpaceX, Neuralink, and xAI
- Giorgia Meloni, Prime Minister of Italy
- Ursula von der Leyen, President of the European Commission
- Sam Altman, CEO of OpenAI
- Nick Clegg, former British politician and president of global affairs at Meta Platforms
- Mustafa Suleyman, co-founder of Google DeepMind
- Michelle Donelan, UK secretary of state for Science, Innovation and Technology
- Věra Jourová, the European Commission's vice-president for Values and Transparency
- Gina Raimondo, United States secretary of commerce
- Wu Zhaohui, Chinese vice-minister of science and technology
The presence of high-level political figures from both Western and Eastern powers underscored the global significance of the event. The participation of industry leaders like Sam Altman and Mustafa Suleyman highlighted the close relationship between AI developers and policymakers, as well as the growing influence of companies such as Anthropic and Google DeepMind in shaping the regulatory landscape.
Global AI Summit series
The Bletchley Park summit was intended to be the first in a series of international meetings on AI safety. The declaration included a commitment to hold future summits, with the aim of building on the initial agreements and addressing emerging challenges. The series was designed to be inclusive, bringing together not only governments but also technology companies, academia, and civil society.
The follow-up summit was scheduled to be held in South Korea in 2024, with a third planned for France. These subsequent meetings were expected to focus on more detailed regulatory proposals, including standards for AI testing and evaluation, as well as mechanisms for international cooperation on enforcement.
The summit series reflected a broader trend toward formalizing AI governance, as countries around the world grappled with the rapid pace of AI development. The discussions at Bletchley Park were informed by research from institutions like MIT CSAIL, Stanford AI Lab, and BAIR (Berkeley AI Research), as well as by the work of prominent AI researchers such as Samy Bengio and Melanie Mitchell, who had called for greater attention to safety and ethics.
Significance and legacy
The Bletchley Park AI Safety Summit was widely regarded as a landmark event in the history of AI governance. It was the first time that a broad coalition of nations, including both democratic and authoritarian governments, had agreed on a shared framework for addressing AI risks. The summit also helped to elevate the issue of AI safety in public discourse, prompting increased media coverage and public debate.
Critics, however, noted that the summit's outcomes were largely symbolic, with the Bletchley Declaration lacking binding commitments or specific enforcement mechanisms. Some researchers argued that the focus on existential risks from frontier AI overshadowed more immediate concerns, such as algorithmic bias, job displacement, and the concentration of power in a few large technology companies.
Despite these criticisms, the summit succeeded in establishing a diplomatic channel for AI governance, which was seen as essential given the global nature of AI development and deployment. The event also highlighted the role of the UK as a convening power in technology policy, building on its strengths in research and innovation.
The legacy of the summit is likely to be measured by the effectiveness of the Global AI Summit series in translating declarations into concrete action. As of 2024, the initial agreements had not yet led to major regulatory changes, but the ongoing dialogue was seen as a positive step toward managing the risks of AI while harnessing its benefits.