KEEP ADVANCED AI UNDER HUMAN CONTROL

To governments, international institutions, and technology companies:

We, the undersigned, call on governments, international institutions, and leading technology companies to take urgent action to ensure that increasingly powerful artificial intelligence systems remain under meaningful human control.

AI systems are becoming increasingly capable of planning, using tools, operating autonomously, and interacting with other systems. As these capabilities advance, the consequences of inadequate safeguards could extend far beyond individual companies or countries.

This is not simply a question of technological progress. It is a strategic issue affecting cybersecurity, economic stability, national security, and the future ability of societies to make decisions about their own affairs.

At the same time, governments and companies are competing intensely to develop and deploy increasingly powerful AI. This creates a difficult incentive: even when organisations recognise serious risks, competitive pressure can make it difficult for any single actor to slow down or invest sufficiently in safety.

We believe no country or company should have to choose between technological leadership and responsible AI development.

WE THEREFORE CALL FOR:

• Independent safety standards for advanced and highly autonomous AI systems, including rigorous testing, containment, monitoring, and auditing.

• International cooperation on AI safety, so that responsible safeguards do not become a competitive disadvantage.

• Greater investment in AI alignment and control research, including methods for monitoring and supervising increasingly autonomous systems.

• Clear thresholds for deployment, requiring stronger safeguards as AI capabilities and levels of autonomy increase.

• Effective incident reporting and information-sharing mechanisms so that dangerous failures and unexpected behaviours are identified and addressed before they become systemic.

• Meaningful human oversight over AI systems capable of taking consequential actions in the real world.

• International dialogue and coordination to address risks that no individual company or government can manage alone.

We are not claiming that catastrophic outcomes are inevitable. We are asking decision-makers to recognise that the consequences of losing control over highly capable autonomous systems could be extraordinarily difficult to reverse.

The opportunity to establish effective safeguards exists now, while humans still have the ability to set the rules.

We call on the world's governments, international institutions, and technology leaders to act before the capabilities of advanced AI outpace our ability to govern and control them.

The development of powerful AI must remain a human-directed process — not a race in which humanity itself becomes the risk.

Please sign this petition and call on decision-makers to make AI safety, oversight, and international coordination a priority.