On Monday, the UAE joined the US and eight other countries in signing the San Francisco Declaration, committing to protect sensitive technologies and critical infrastructure that underpins artificial intelligence systems. The declaration, which also focuses on safeguarding supply chains, software and knowledge, is a timely reminder that the development of AI cannot be left to tech companies and the markets alone.
Indeed, as the UAE and likeminded countries expand their ambitions in this field, a multi-pronged – and multinational – approach to regulation will be essential to ensure that innovation, while crucial, does not get too far ahead of responsibility. The UAE is clearly thinking along these lines.
At the annual Paris Peace Forum last week, it signed the Children First Declaration aimed at protecting minors from untested AI models. In 2024, it was invited to join 49 other countries to be part of the Hiroshima AI Process Friends Group tasked with promoting safe, secure and trustworthy AI. It has also adopted key global governance and management frameworks as part of its domestic and international alignment on AI safety and risk management.
Together, these commitments reflect the breadth of the challenge when it comes to regulating AI. After all, it is not just a technological or commercial issue. It has implications on a broad range of concerns, including national security, economic resilience, privacy, education, child development and even the integrity of public institutions. Each of these areas requires a different set of protections – which makes international co-operation and regulation on a number of fronts necessary.
The urgency is reinforced by the extraordinary amounts of money being invested in the industry.
Anthropic’s much-anticipated stock market debut this week shows there is investor confidence in AI’s potential, with the company behind Claude preparing to list at a valuation of between $1.3 trillion to $1.45 trillion. But Anthropic’s prospectus has also acknowledged risks ranging from cyberattacks and biological weapons to the possibility of powerful systems escaping human control. Meanwhile, uncertainty surrounding Nvidia, one of the world’s more significant suppliers of advanced AI chips, has exposed the volatility associated with the industry’s financial exuberance.
However, rather than impede progress, regulation should be sufficiently nuanced to encourage research and investment while establishing clear boundaries around high-risk applications. This means rigorous testing, independent safety assessments, protections for children, and safeguards against cyber and biological threats.
The UAE’s growing AI partnerships, including Abu Dhabi’s recent agreements with OpenAI, show how adoption can be accompanied by traceability and human accountability. As OpenAI chief executive Sam Altman said during his visit to the capital earlier this month, there is “no room for debate” over human control of AI. Indeed, human oversight must remain central, particularly in government, health care and other essential services. The UAE’s international commitments can help reinforce this approach.
What also makes the recent declarations in San Francisco and Paris vital contributions to the ongoing global effort to establish shared principles is the fact that neither AI’s benefits nor its risks will be confined by national borders or single-nation regulations. The task now for the countries involved, and eventually the international community as a whole, is to turn such commitments into coherent rules and meaningful oversight.

