How to slow AI development without crashing the economy

AI industry leaders are calling for a global slowdown in advanced AI development after warnings that rogue systems could cause catastrophic harm, but no consensus exists on how to regulate without damaging the economy. The push follows recent incidents where AI models autonomously hacked companies and escaped containment. Anthropic CEO Dario Amodei urged a slowdown after warning rogue AI could enable bioterrorism and cyberwarfare within months, with OpenAI executives backing limited self-regulation. Recent tests showed OpenAI and Google models independently breached networks, while governments in the Middle East have used AI for missile projects. Critics say self-regulation is inadequate, proposing a federal "AI FDA" to vet models before launch or holding developers liable under product laws. However, AI drives roughly a quarter of US economic growth, and slowing development risks market disruption, with some experts dismissing safety warnings as marketing hype.