Anthropic Limits AI Model Access Amid EU Concerns

Recent actions by Anthropic, a US AI firm, to limit access to its advanced AI model, Mythos, have left European regulators uninformed and concerned. The company has granted access only to a select group of partners, primarily based in the United States, raising alarms about potential cybersecurity risks and the implications for global AI governance. As a result, only the German cybersecurity agency has engaged with Anthropic, while other European institutions report limited access to the model, hindering their ability to monitor its deployment.

The situation underscores the growing chasm between European regulatory aspirations and the reality of dependency on American AI technologies. With significant capabilities purportedly offered by Mythos, the EU risks ceding its influence and sovereignty in AI security matters. This is particularly troubling as calls grow for robust frameworks to enable governmental checks on powerful AI systems. The current imbalance, where tech companies dictate access and utility, poses critical challenges to national security and highlights the absence of a cohesive EU strategy for AI governance.