While Wall Street celebrated another strong session for artificial intelligence (AI) stocks, including SpaceX’s impressive post-IPO surge, a significant regulatory cloud has suddenly appeared over the sector. At the centre of this emerging challenge is Anthropic, an AI model maker known for developing advanced AI systems and large language models. Its chief executive, Dario Amodei, has frequently positioned himself as a deep-thinking philosopher on AI regulation.
The US government has now placed unprecedented export controls on Anthropic’s latest AI model, Fable. This consumer-friendly version of its earlier Mythos model was pulled from the market after Amazon chief executive Andy Jassy alerted top White House officials to a potential security risk. Amazon researchers identified a “jailbreak” in Fable, a method to bypass security guardrails. Anthropic was given less than 90 minutes to comply with an order barring foreign nationals, though the company disputes the alleged vulnerability and is working to restore its model and government relationship.
This extraordinary government intervention has sent shockwaves through the AI world, prompting a swift reality check for an industry accustomed to rapid growth. Ironically, this development follows Amodei’s own recent essays advocating for government power to block AI model deployments deemed risky. The export ban raises crucial questions for the entire AI industry regarding increased regulatory burdens, the deployment of models beyond American borders, and the financial growth trajectory of companies like Anthropic. Generating revenue from a controlled product presents an immediate hurdle, especially amid a potential AI price war and the massive investments, like Nvidia’s recent US$25 billion bond sale, riding on the sector’s sustained expansion.
