UK leads on newly signed global AI cybersecurity guidelines

The UK has led the development of the first-ever globally agreed artificial intelligence cyber-security guidelines.

A total of 18 countries have signed the Guidelines for Secure AI System Development, which mark a “key milestone” in the advancement of the technology’s capabilities, according to Cybersecurity and Infrastructure Security Agency (CISA) director Jen Easterly.  

Among signatory countries are Australia, Canada, Germany, Nigeria and the Republic of Korea.   

The announcement comes amid a growing fear across businesses on the rapid development of the technology, with 10 per cent of Scottish companies having already banned employees from using generative AI, according to a recent labour report. 

“As nations and organisations embrace the transformative power of AI, this international collaboration underscores the global dedication to fostering transparency, accountability, and secure practices”, Easterly said. 

Developed by GCHQ’s National Cyber Security Centre, the recommendations are divided into four areas – secure design, secure development, secure deployment, and secure operation and maintenance – and include a list of behaviours businesses can adopt to enhance security. 

Written in partnership with CISA – a US agency – and industry representatives, the guidelines are based on a secure-by-design approach.  

In other words, a method that outlines cyber security as a pre-requisite for the development of AI safety systems. 

Among the 21 other international agencies that co-operated in the creation of the guidelines are G7 and the Global South member nations. 

Alejandro Mayorkas, secretary of the US Department of Homeland Security, said: “We are at an inflexion point in the development of AI, which may well be the most consequential technology of our time. Cyber security is key to building AI systems that are safe, secure, and trustworthy. 

“By integrating ‘secure-by-design’ principles, these guidelines represent a historic agreement that developers must invest in, protecting customers at each step of a system’s design and development.” 

The announcement adds to the UK’s pledge to lead on the design of the international regulation of AI. 

The commitment was showcased earlier this month at the AI Safety Summit at Bletchley Park, where 28 countries signed a world-first declaration to deal with AI.