The Commerce Department confirmed on Friday that it has rescinded the export controls placed on Anthropic’s flagship AI models, Mythos 5 and Fable 5. In a brief statement, the agency said the ban, which had forced the startup to block all external access, would be lifted effective July 1. Anthropic promptly announced on its X account that it would begin restoring access to its customers the following day.
The restriction, first imposed in late June, required Anthropic to suspend any use of the models by foreign nationals, even when those individuals were employed by the company in the United States. The order came after the department received alerts from major tech firms, including Amazon, that the models were susceptible to jailbreak techniques that could enable malicious actors to bypass built‑in safety filters.
In response, Anthropic halted all traffic to Mythos and Fable, citing compliance with the government directive. The company also published a detailed list of safeguards it had put in place, ranging from automated monitoring to manual review processes, to address the highlighted vulnerabilities.
Earlier this month, the department granted Anthropic limited permission to redeploy Mythos 5 for a select group of roughly 100 U.S. organizations tasked with operating and defending critical infrastructure. Those partners, part of Anthropic’s Project Glasswing, received access to the model’s advanced cybersecurity capabilities under strict oversight.
Now, Commerce Secretary Howard Lutnick, in a letter to Anthropic, indicated that the broader ban on both Mythos 5 and Fable 5 would be lifted once the company demonstrated a proactive stance on risk detection and mitigation. "We are lifting the ban on Fable after the company agreed to proactively detect and address security risks associated with the models," Lutnick wrote.
Anthropic’s CEO, Dario Amodei, praised the decision, saying the company had worked closely with regulators to ensure the models could be used safely. "We appreciate the patience of our users and the collaborative effort with the Department of Commerce to get these tools back online," he said in a follow‑up post.
The reversal marks a significant shift in U.S. policy toward emerging AI technologies. While regulators remain vigilant about the potential for misuse, the move signals a willingness to balance national‑security concerns with the commercial and research benefits that powerful AI models can deliver.
Industry observers note that the episode underscores the growing tension between rapid AI innovation and the need for robust safeguards. Companies developing large‑scale models now face heightened scrutiny, and the Anthropic case may serve as a template for future engagements between the tech sector and federal agencies.
Anthropic plans to monitor usage closely and will provide regular updates to the Commerce Department. The company also pledged to continue refining its safety layers, aiming to prevent the kinds of jailbreak attempts that triggered the original export controls.
Este artigo foi escrito com a assistência de IA.
News Factory APP - notícias agênticas para impulsionar seu SEO e AEO.