Source: Fortune
Summary
Anthropic has restored global access to its AI model Fable 5, two weeks after the Trump administration imposed export controls citing security concerns. The company announced that the controls had been relaxed, and US Commerce Secretary Howard Lutnick confirmed that the government had worked with Anthropic to ensure the model aligns with US interests. The restrictions were initially imposed due to concerns that hostile actors could “jailbreak” the models and bypass built-in guardrails. Anthropic will work closely with the White House and other companies to develop a shared industry framework for assessing potential vulnerabilities in frontier models.
Our Reading
The numbers tell one story.
Anthropic’s announcement suggests a thawing in its relationship with the Trump administration. The company will scale up government collaboration and work with Amazon, Microsoft, and Google to set industry standards for assessing potential vulnerabilities in frontier models. Anthropic’s willingness to work closely with the government may be driven by its reliance on federal contracts and its upcoming IPO plans. The company’s valuation of $1 trillion and its highly anticipated IPO make it a high-stakes player in the AI industry.
The strategy enters a familiar phase: Anthropic is walking a fine line between appeasing the government and maintaining its independence. The company’s decision to prerelease frontier models for federal review and implement dedicated research teams to ensure its models align with government priorities suggests a balancing act. As Anthropic navigates this delicate relationship, its IPO plans and valuation hang in the balance.
Author: Evan Null
