United States Commerce Secretary Howard Lutnick has ordered artificial intelligence company Anthropic to suspend exports of its latest Mythos and Fable AI models worldwide amid concerns that the technology could potentially be diverted for use by military intelligence entities in countries identified as strategic concerns, including China and Russia. According to a letter sent to Anthropic Chief Executive Officer Dario Amodei and reviewed by Reuters, the directive instructed the San Francisco based company to halt access to the models for foreign destinations and foreign nationals regardless of location. The action followed concerns raised by U.S. officials regarding the possible misuse of advanced artificial intelligence systems in activities that may affect national security. The decision also prompted discussions between Anthropic representatives and U.S. Department of Commerce officials as both sides sought a path forward regarding the deployment of the company’s latest AI models.
According to reports, senior technical staff from Anthropic met with Department of Commerce officials in Washington to discuss potential solutions after the company disabled access to the models globally following the government directive. Officials reportedly expressed concerns that security safeguards integrated into Anthropic’s recently introduced Fable 5 model may be bypassed through a process commonly referred to as “jailbreaking,” allowing users to identify software vulnerabilities. Anthropic stated in a company blog post that while bypass methods had been identified, they exposed only limited or minor vulnerabilities that other publicly available artificial intelligence models are already capable of identifying. Sources familiar with the discussions stated that Commerce Secretary Howard Lutnick has remained directly involved through regular communication with Anthropic executives as efforts continue to reach an agreement balancing technological access with national security priorities. Both Lutnick and Anthropic Chief Executive Officer Dario Amodei are also expected to attend G7 meetings in Evian les Bains, France, where further discussions may take place alongside ongoing negotiations.
The dispute follows earlier tensions between Anthropic and Trump administration officials after the company reportedly declined requests to allow its artificial intelligence systems to be used for domestic surveillance purposes and fully autonomous weapons systems. According to reports, the disagreement contributed to Anthropic being added to a national security blacklist earlier this year. Anthropic had previously delayed broader release of its Mythos model after raising internal concerns regarding advanced cybersecurity capabilities associated with the technology. On June 9, the company introduced a public version called Fable 5, which included cybersecurity protections designed to limit misuse. According to a source close to the company, Anthropic worked alongside government officials to evaluate the model prior to launch and had reportedly received approval for deployment before the latest restrictions were introduced.
The action taken by Department of Commerce reportedly marks the first time authorities under Export Control Reform Act of 2018 have been used to restrict an artificial intelligence model categorized as an emerging technology linked to national security considerations. The directive requires export licences for transferring access to foreign destinations or foreign nationals located within United States and warns of potential criminal and civil penalties for non compliance. However, export control specialists have questioned the legal basis for the restriction, arguing that artificial intelligence models are generally deployed through remote access rather than conventional exports covered by regulatory frameworks. More than 80 cybersecurity executives and experts reportedly signed an open letter addressed to Howard Lutnick and National Cyber Director Sean Cairncross supporting Anthropic’s position and urging reconsideration of the restrictions. Executives associated with major technology companies, including Nvidia and Adobe, were among those backing removal of the measures while discussions between Anthropic and U.S. officials continue.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
