Amazon CEO Andy Jassy Raises Alarm Over Anthropic AI Models, Triggering Global Access Block—Report
A major development in the global artificial intelligence industry has emerged after reports that Anthropic’s advanced AI models were restricted following security concerns raised by Amazon CEO Andy Jassy. According to the Wall Street Journal, Jassy’s warnings to US officials played a key role in triggering a government response that ultimately led to a global access block on some of Anthropic’s most powerful systems.
The report claims that Jassy directly contacted senior US officials, including Treasury Secretary Scott Bessent, after Amazon researchers allegedly discovered that Anthropic’s “Fable 5” model could be manipulated through specific prompts. These prompts, according to the findings, could potentially generate information useful for cyberattacks—something the model was designed to prevent. The concerns quickly escalated within US security circles.
Following the alert, White House officials and security experts reportedly conducted further assessments and confirmed the need for urgent action. Anthropic was asked either to fix the vulnerability or take the model offline entirely. However, after internal deliberations, US authorities concluded that the safest option was to block access to the model for foreign governments, companies, and individuals, prioritizing national security over global availability.
The decision was ultimately approved by President Trump, despite internal concerns that such restrictions could slow down AI innovation. A follow-up discussion between US officials and Anthropic CEO Dario Amodei reportedly failed to ease tensions, with some officials expressing doubts about the company’s willingness to fully cooperate with government security teams. Anthropic, however, pushed back on the claims, arguing that the issue involved a limited “jailbreak” vulnerability and not a widespread security flaw.
In response to the directive, Anthropic reportedly disabled global access to its “Mythos” and “Fable” models to comply with government expectations. The move has raised wider concerns across the tech industry, especially among international developers who rely on such tools for cybersecurity research and software development. The controversy now highlights a growing tension between rapid AI innovation, national security interests, and global access to cutting-edge technology.
source: nairametrics