Key facts
- Global governments are struggling to find a path forward on AI without US buy-in.
- US President Donald Trump rejected any "globalist scheme of control" on AI.
- OpenAI CEO Sam Altman called for countries to come together for shared interests in the face of new technology.
- Anthropic CEO Dario Amodei pushed for "common global standards" for testing AI models.
- The US and China did not sign a declaration by 22 countries pushing for a global AI supervisory regime.
- China supports the formation of a "consensus-based global governance framework" for AI.
Global efforts to establish international protections and guardrails for artificial intelligence are facing significant hurdles due to the United States' reluctance to participate in what President Donald Trump has described as a "globalist scheme of control." World leaders and tech executives gathered at the U.N. General Assembly this week to advocate for a unified approach to AI safety, particularly in light of recent cyberattacks.
Despite the US stance, many nations are pushing forward. The UN Security Council recently hosted leaders from top US AI companies, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, who voiced urgent concerns about AI security and advocated for international cooperation and common global standards for AI model testing. Amodei warned that poorly managed AI could pose a risk to humanity.
Earlier in the week, 22 countries, spearheaded by Finland and Norway, adopted a declaration calling for a global supervisory regime for AI safety and security, emphasizing that AI must remain under human control. However, neither the United States nor China endorsed this declaration. The majority of leading AI companies are based in these two nations.
James Lewis, a former State Department official, noted that without the participation of major AI players like the US and China, any agreements on AI weaponization would be meaningless. However, China's permanent representative to the UN, Fu Cong, expressed support for a consensus-based global governance framework. UN Secretary-General António Guterres highlighted the responsibility of nations with the greatest AI capabilities.
While a recent POLITICO poll indicated that a majority of adults in the US, Canada, and Europe favor pausing AI development due to existential fears, many European officials expressed frustration with the US rejection, though few are willing to openly criticize Trump. France, meanwhile, plans to actively participate in upcoming UN cybersecurity initiatives and advocates for sovereign states to collaborate on open-source AI models.
The UN has been attempting to establish itself as a venue for AI discussions, with the Independent International Scientific Panel on AI releasing a preliminary report on AI risks. The UN's Global Dialogue on AI Governance is scheduled for next May. However, the Trump administration's general downgrading of support for global cooperation, coupled with the lengthy process of creating international treaties, presents a challenging path forward for AI policy.