Former Heads of State Issue Eight AI Principles for Global Safety

A coalition of former heads of state has published a set of principles aimed at reshaping how governments manage artificial intelligence. The group argues that commercial interests and geopolitical rivalry have effectively sidelined the international cooperation needed to mitigate the technology's risks. They contend that a small number of private companies, largely based in the United States and China, are currently shaping the trajectory of one of the century's most disruptive transformations without sufficient oversight.

The document, released via The Elders.org, explicitly rejects the notion that AI is an uncontrollable force of nature. Instead, it frames the technology as a market product built by companies, subject to the same legal and ethical standards as other industries. The authors warn that the harms of AI may fall disproportionately on populations with the least power to influence its development, calling for a governance model that places human well-being ahead of profit and national power.

Governments Must Assume Primary Safety Roles

The core argument of the principles is that the state cannot outsource its duty to protect citizens to private developers. The text emphasizes that safety is a government responsibility, not a corporate one. It calls for leaders to enforce existing laws, draft new regulations where gaps exist, and break up concentrated power structures that have grown excessive. The authors note that setting strict rules does not mean falling behind in innovation, but rather ensuring that progress is safe and beneficial for the broader public.

A significant portion of the guidance focuses on the need for international cooperation. The authors argue that no single nation can manage the cross-border risks of AI alone. They propose that international agreements should encourage responsible development while still allowing for commercial competition. The goal, they state, is not to slow down technological progress but to create a framework where AI is trustworthy and its benefits are widely shared.

Transparency and Human Oversight Are Required

The principles place heavy emphasis on trust and accountability. Leaders are urged to communicate the truth about AI's capabilities and limitations to their citizens. This includes disclosing known risks, potential dependencies, and the likely distribution of economic gains and losses. The text warns against allowing core state functions, such as public services and essential infrastructure, to be outsourced to systems that the public cannot examine or hold accountable. Independence from the industry is framed as a prerequisite for effective governance.

Restraint is another central theme, particularly regarding military applications. The authors insist that the legitimate use of force must remain a human responsibility. They explicitly call for the rejection of autonomous weapons that select their own targets and commercial systems that influence how targets are chosen. This stance reflects a broader insistence that AI must serve human beings rather than threaten their rights or existence.

Inclusive Decision-Making Expands Beyond Tech Hubs

The final principle addresses solidarity, urging leaders to involve those affected by AI in the decision-making process. This includes communities bearing the environmental costs of data centers and countries that have been excluded from the current geopolitical dynamics of AI development. The authors argue that those who suffer from or consume the technology should not be passive recipients of terms set elsewhere. By broadening the circle of stakeholders, they hope to create a more globally inclusive and long-term view of AI governance.