Anthropic CEO Dario Amodei says AI industry needs to give safety measures time to catch up

Anthropic CEO Dario Amodei’s call for a slowdown in AI development, citing a six to twelve month window for potential internet-scale agent control, reflects a growing internal industry debate. For Asia, this discussion is critical as regional governments and tech firms like Tencent and Baidu are heavily investing in AI. A global consensus on AI safety standards, or lack thereof, directly impacts how quickly Asian companies can deploy new models and applications without facing regulatory hurdles or public backlash. The risk is that a fragmented approach to safety could create uneven playing fields, potentially hindering innovation in markets that adopt stricter controls early on. Our view is that Amodei’s warning, while stark, underscores the need for proactive policy development in Asia. Countries like Singapore and South Korea, which aim to be AI leaders, must balance rapid innovation with robust safety frameworks. The alternative is a reactive regulatory environment that could stifle growth or, worse, expose critical infrastructure to unforeseen AI-driven risks. The coming year will show whether these warnings translate into concrete, collaborative action across the region.