US, UK enter agreement to develop AI safety standards

The US and the UK signed a Memorandum of Understanding (MOU) to make standards for AI Safety connected April 1.

The statement aims to make and iterate “robust suites of evaluations” for AI models, systems, and agents done coordinated technological approaches.

The 2 countries volition specifically behaviour astatine slightest 1 associated investigating workout connected a publically accessible AI model. They besides mean to speech unit betwixt their antecedently established AI Safety Institutes. The concern takes effect immediately.

The US Department of Commerce and the UK’s Department for Science, Innovation and Technology announced the concern successful identical property releases.

Safe AI development

Commenting connected the caller planetary agreement, US Commerce Secretary Gina Raimondo called AI the “defining exertion of our generation” and said that the concern volition code each risks related to the technology, from nationalist information risks to societal risks.

Raimondo described the partnership’s proactive nature, stating:

“Our concern makes wide that we aren’t moving distant from these concerns – we’re moving astatine them.”

UK Technology Secretary Michelle Donelan called the statement a “landmark moment” and called the pursuit of harmless AI improvement a planetary issue.

In a separate interview with the Financial Times, Donelan emphasized the urgency of the partnership, considering the upcoming merchandise of caller AI models implicit the adjacent year. She added that these models’ afloat capabilities are unknown, but they are expected to beryllium “game-changing.”

Donelan highlighted that galore apical AI companies are US-based, which necessitates accusation speech with the government.

Other AI developments

The existent statement builds connected commitments to AI made during the UK’s AI Safety Summit, which was held successful November 2023. The broader committedness to AI information is portion of the Bletchley Declaration, a argumentation insubstantial from countries that attended the summit.

The quality comes soon aft the completion of goals successful the Biden Administration’s enforcement bid successful the US, which culminated successful orders to name main AI officers crossed each national agencies connected March 28.

The UK acceptable retired a section attack to regularisation successful February.

The station US, UK participate statement to make AI information standards appeared archetypal connected CryptoSlate.