British officials thought they had achieved a diplomatic victory while sitting in a specially constructed hut located on the property of the historic Bletchley Park estate.
On stage in front of them was the UK’s technology secretary, Michelle Donelan, and behind her were high-level representatives from the US and China, together for the first time to discuss the international regulation of artificial intelligence.
Moreover, both nations were part of the 28 signatories who signed the “Bletchley declaration”, a pact to collaborate on safety protocols that could potentially stop AI systems from becoming a threat to humanity.
On Thursday, Rishi Sunak stated that there were those who believed China should not be invited and others who thought it would be impossible to reach an agreement with them. However, both of these beliefs were proven wrong. A comprehensive approach to ensuring AI safety must involve collaboration with all of the leading AI countries, and this has been demonstrated through the signing of the Bletchley Park communique by all of them.
The delegates were happy that China was participating actively in the discussions, despite not being invited to the more sensitive talks on the second day.
The British team, headed by experienced mediator Jonathan Black and tech investor Matt Clifford, worked tirelessly to convince Beijing to participate in the conference and agree to the statement. This included securing permission for officials to be present for both days of the event.
Despite the friendly diplomatic gestures, there are indications that the leading global powers are competing for dominance. Each desires to take the lead in establishing regulations for the advancement of AI, recognizing that this could also give an advantage to their own domestic agendas and businesses.
Certain officials in the United Kingdom had anticipated receiving global recognition for utilizing their AI taskforce as a central location for trialing new systems before making them available to the public. However, Gina Raimondo, the Secretary of Commerce in the United States, revealed that their government will establish their own institute for this purpose, and the two entities will collaborate closely.
Prior to the commencement of the summit, Joe Biden issued an executive order on AI, which according to British diplomats, aided in directing the discussions at Bletchley.
The European Union is currently making significant progress in creating new regulations for AI. Věra Jourová, the vice-president for values and transparency at the European Commission, stated that the UK is lagging behind due to their own choice. On the other hand, Chancellor Rishi Sunak believes it is premature to implement legislation. He argued, “How can we create laws for something that we do not fully comprehend yet?” during his remarks last week.
There are conflicting views on how to handle open-source AI, which allows for the use and development of new models by anyone. The British and American governments are concerned that this could give individuals the ability to cause significant social harm, while the French are more optimistic about its potential. Yann LeCun, a prominent French AI innovator and chief AI scientist at Meta (Facebook’s parent company), strongly advocates for open-source.
Tino Cuellar, president of the Carnegie Endowment for International Peace, attended the summit and stated that there are countries, such as those in the European Union and France, as well as some in the global south, who are generally interested in the accessibility of this technology to multiple countries.
However, representatives from multiple nations are convinced that they are establishing the foundation for a system that can connect different types of national regulations.
On Thursday, Sunak declared that nations worldwide have come to an agreement to establish an international group of experts to assess the capabilities of AI, similar to the Intergovernmental Panel on Climate Change. Additionally, companies have consented to government review of their most cutting-edge AI models for safety, although this will not be mandatory.
There are different opinions on the new system, with some drawing parallels to air traffic control regulations. In this scenario, there are minimum standards agreed upon globally, but individual countries have the freedom to impose stricter standards. Alternatively, some suggest modeling the system after the drug approvals process, where certain countries recognize the decisions made by other countries’ regulators.
On Thursday, attendees at the primary summit location noted that it was only half occupied. Many prominent guests, such as Canadian Prime Minister Justin Trudeau, opted to participate remotely.
However, numerous individuals in the field were pleased to witness global leaders openly addressing the dangers that they have been cautioning about behind closed doors for a significant amount of time.
“I have spent approximately a decade advocating for a global summit on the safety of advanced AI. Witnessing this event come to fruition, and seeing it executed so successfully, was an incredibly emotional experience,” stated Max Tegmark, a prominent leader in the movement to pause the development of advanced AI.