The United States and the United Kingdom is collaborating to conduct safety tests and risk assessment on AI models amid dangers posed by the technology, the countries' tech departments announced on Monday.
US Commerce Sec. Gina Raimondo and UK Technology Sec. Michelle Donelan said the two departments will be partnering to "accelerate and rapidly iterate robust suites of evaluations for AI models, systems, and agents."
According to the secretaries, the first step in the partnership will be assessing publicly available AI models on their impact on national security and "our broader society."
The partnership is part of the two nations' commitment during the AI Safety Summit five months ago.
The US and UK earlier opened their respective AI Safety Institutes to focus on the effects of the rapidly increasing AI development in their countries.
US, UK Prepares for 2024 Elections
With both countries heading for election polls, the collaboration to safety test AI marks the US and the UK's preparation to address dangers and risks emerging from the technology.
Several experts have already warned the governments to buckle for a surge in disinformation and cyberattacks as the election period inches close.
Ahead of government bodies intervening in their operations, AI companies like OpenAI and Google have started imposing stricter policies to prevent people from abusing their products to disrupt elections.
However, safety tests revealed that some prompt adjustments still result in the AI generating deepfakes of political figures and political misinformation.
Standardized AI Regulations Remain a Far Goal for US, UK
Despite the increased efforts in addressing issues posed by AI, both countries have yet to implement standardized regulations to govern AI development, distribution, and application.
Several legislations aimed to regulate AI have been filed in the Parliament, although the government hints at seeking a more flexible approach to governing the technology.
The US, on the other hand, balances proposed regulations on AI on an executive order issued last October. A proper AI law has yet to pass Congress despite the current administration's push on the trend to revive its tech industries.
As of writing, only the European Union has adopted an actual AI law to govern risks and approach to the technology within bloc members.
