# Anthropic CEO Proposes AI Development Slowdown Amid China Dilemma Dario Amodei urges coordinated global limits on AI progress to prevent catastrophic risks. - **Source:** [US Top News and Analysis](https://brifing.si/source/635540e9-943e-450d-ae52-ae0bfe3b67eb) - **Published:** 2026-09-13 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/7690287d-d02b-4a3f-8a9a-b0bf2b3d8dff ## The China Dilemma Anthropic CEO Dario Amodei identifies China as the toughest challenge in implementing his proposed AI slowdown, citing military advantages as key incentives for nations to push ahead. He admits uncertainty about achieving international cooperation but emphasizes the necessity of trying. This comes as AI's potential for catastrophic harm becomes increasingly apparent, with industry researchers issuing stark warnings last week. ## Three-Step Safety Plan Amodei proposes a comprehensive approach: 1) Third-party safety audits with employee-level access, which Anthropic has already implemented unilaterally. 2) Coordination among democratic nations' AI companies to establish common standards. 3) Engagement between democratic and authoritarian governments. The plan aims to balance safety with maintaining US leadership in AI, addressing both commercial and geopolitical concerns. ## Industry Reactions The proposal received mixed but notable support. OpenAI's Sam Altman agreed on pacing development despite past tensions with Amodei. Elon Musk, previously critical of Anthropic, endorsed peer review after SpaceX's $1.25B/month compute deal with the company. Google DeepMind's Demis Hassabis called it 'the right path forward,' while Meta's AI chief Alexandr Wang emphasized scaling alignment efforts. ## Researcher Resignation Jacob Coxon, an Anthropic researcher, publicly resigned over concerns that Anthropic and OpenAI are 'gambling with our lives.' His departure highlights growing ethical concerns within AI development teams about the rapid pace of advancement without adequate safeguards. This follows increasing public backlash against AI companies regarding data center impacts and calls for regulation. ## IPO Considerations The debate occurs as both Anthropic and OpenAI prepare for potential historic IPOs, though neither has set dates. OpenAI's Altman stated 2024 is unlikely due to safety concerns, calling current timing 'ill-advised.' Meanwhile, industry observers speculate Anthropic might go public first. These financial moves intersect with safety discussions, as public offerings could increase pressure for rapid advancement. ## Alignment Challenges Meta's AI chief Wang noted alignment - ensuring AI matches human values - is becoming a 'gating factor' as models grow more powerful. This technical challenge compounds the geopolitical ones, requiring significant resources as models approach superintelligence levels. The industry faces balancing rapid scaling with increasingly complex alignment requirements. ## Critics' Perspective David Sacks, former Trump administration AI official, supported the slowdown but criticized perceived hypocrisy. He argued companies shouldn't demand regulatory frameworks while claiming existential risks, calling it potential 'blackmail.' Sacks noted liability concerns may motivate slowdowns more than altruism, as products enabling cyberattacks could expose firms to massive legal risks. ## Global Implications The proposal underscores AI's dual-use dilemma: technological leadership versus existential risk. With China unlikely to voluntarily slow down, democratic nations face coordinating among themselves while engaging adversaries. The outcome could shape not just AI safety but global power dynamics, making this one of the most consequential industrial policy challenges of our time. ## Sources covering this story - https://www.cnbc.com/2026/09/13/china-dilemma-ai-slowdown-anthropic.html — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._