The rapid acceleration of artificial intelligence poses a fundamental challenge to humanity and requires urgent regulatory action, according to the chief executive of AI company Anthropic. He said technological progress is advancing faster than legal systems and social institutions can adapt, creating risks that extend beyond economics into political power and civil liberties. In a recent essay, the executive argued that artificial intelligence is entering a stage of development where its capabilities could soon surpass those of top human experts across a wide range of disciplines. He warned that systems operating at this level, particularly when deployed at scale, could dramatically alter how governments, organisations, and societies function. Without clear safeguards, he said, AI could become a tool that concentrates power rather than distributing opportunity, increasing the risk of misuse both state and non state actors at a time when global governance frameworks remain fragmented.
The executive described the current phase of AI development as a period of technological adolescence, marked rapid capability growth and limited oversight. He suggested that advanced systems working together could perform complex tasks far more efficiently than human teams, influencing areas such as software development, cyber operations, and strategic decision making. This combination of speed, autonomy, and difficulty of control raises concerns about unintended consequences, particularly if deployed without sufficient transparency. He warned that authoritarian governments could use AI to strengthen surveillance and suppress dissent, potentially locking societies into permanent systems of control. According to his assessment, large scale monitoring powered advanced AI could allow authorities to anticipate and neutralise opposition before it emerges, fundamentally altering the balance between citizens and the state.
Beyond state misuse, the executive said democracies, corporations, and technology developers themselves must also confront the risks associated with powerful AI systems. He argued that controlling access to advanced computer chips used to train large models represents one of the most effective safeguards, calling for tighter export controls to prevent misuse. He also urged greater coordination across the technology sector, including transparency requirements that compel companies to explain how their systems are governed and aligned with ethical standards. While acknowledging the scale of the challenge, he expressed cautious optimism that decisive action could mitigate the dangers. He said artificial intelligence remains a transformative opportunity, but warned that its benefits depend on recognising the technology as a civilisational issue requiring global cooperation, restraint, and long term oversight.
