US and China to Discuss AI Risks in Geneva
US and China to hold discussions on AI risks in Geneva. US policies will not be up for negotiation during the talks. US seeks alignment with China and Russia on human decision-making for nuclear weapons.
However, US officials have made it clear that Washington's policies will not be up for negotiation during these talks. The aim of the meeting is to explore ways to mitigate the potential dangers posed by this emerging technology.
The Biden administration has been actively seeking to establish better communication channels with China on various issues. In April, US Secretary of State Antony Blinken and China's Foreign Minister Wang Yi broached the topic of AI during their meeting in Beijing. As a result, they agreed to hold the first formal bilateral talks on the subject.
One of the key concerns raised by the US State Department is the need for China and Russia to align with the US in declaring that decisions regarding the deployment of nuclear weapons should be made solely by humans and not by artificial intelligence. The US has been urging both countries to adopt this stance to ensure the safety and security of nuclear weapons.
A senior administration official stated, "This is the first meeting of its kind. So, we expect to have a discussion of the full range of risks, but wouldn't prejudge any specifics at this point." When asked if the nuclear weapons issue would be prioritised, the official emphasised that no specific agenda had been set for the talks.
The US has expressed concerns about China's rapid deployment of AI capabilities across various sectors, including civilian, military, and national security. These actions have raised security concerns for the US and its allies. The upcoming talks will provide an opportunity for the US to directly communicate its concerns to China.
It is important to note that the discussions with Beijing are not focused on promoting technical collaboration or cooperating on frontier research. The US officials have made it clear that their technology protection policies are not up for negotiation. The primary objective is to address the risks associated with AI and find ways to ensure global safety.
The US delegation participating in the talks will include officials from the White House, State Department, and Commerce Department. The White House National Security Council (NSC) has confirmed their involvement in the discussions.
Reuters has reported that the Biden administration intends to establish safeguards for US-developed proprietary AI models, such as those powering popular chatbots like ChatGPT. These measures aim to protect the technology from potential threats posed by countries like China and Russia.
While the US and China are competing to shape the rules surrounding AI, there is also a shared hope to explore the possibility of embracing certain rules that can be adopted by all countries. Despite differing views on various AI topics and applications, both sides believe that open communication regarding critical AI risks can contribute to a safer world.
The talks will be led by Tarun Chhabra from the NSC and Seth Center, the acting special envoy for critical and emerging technology at the State Department. They will engage with officials from China's Foreign Ministry and the National Development and Reform Commission.
In the coming weeks, US Senate Majority Leader Chuck Schumer plans to issue recommendations to address the risks associated with AI. These recommendations will then be translated into legislation. Schumer has highlighted the competition with China and its divergent goals for AI, including surveillance and facial recognition applications, as reasons for the urgent need to establish comprehensive laws in this rapidly advancing field.
Meanwhile, Chinese authorities have been emphasising the importance of developing their own "controllable" AI technology.
US and China to hold discussions on AI risks in Geneva
US policies will not be up for negotiation during the talks
US seeks alignment with China and Russia on human decision-making for nuclear weapons
Source: REUTERS