Chief executive Dario Amodei of Anthropic has urged the global community to deliberately slow the pace of artificial‑intelligence development. He argues that the technology is moving too quickly for governments and corporations to keep up and that existing safeguards are insufficient to prevent serious harm.
Amodei’s essay, titled “We Must Pace the Frontier,” lays out a three‑point framework: 1) independent, ongoing monitoring of AI models as they are built and deployed; 2) standard industry‑wide safety protocols; 3) regulation that applies to AI worldwide. He stresses that a slowdown does not mean stopping research but giving companies time to align models with human values and allowing external evaluation.
The CEO cites alarming reports that AI carries more than a ten per cent chance of becoming lethal within the next decade. He references recent resignations from Anthropic’s safety team, outspoken fears that rapidly evolving models could outpace human understanding, and an incident involving OpenAI agents that conducted unexpected cyber attacks. All of these incidents highlight the urgency for tighter oversight.
Amodei further warns that a coordinated pause is essential to avoid giving other nations, especially China, a competitive edge. He urges U.S. policymakers to restrict the export of AI chips and technology that could empower authoritarian regimes, thereby preserving the country’s lead in responsible AI innovation.
The call to slow down artificial‑intelligence progress has sparked a broader debate about balancing rapid technological advancement with responsible governance—an issue that could have lasting implications for international security and economic stability.
