Sarah Heck
Coverage of Sarah Heck in the Nexus archive.
- ‘Anthropic doesn’t care about Europe’ — EU officials peeved after AI giant sends junior staffer to testify about safety
European policymakers criticized Anthropic for sending a junior technical employee, Donny Greenberg, instead of a senior official to address concerns about AI models Mythos and Fable, which pose cybersecurity risks. EU officials accused the company of dismissing Europe’s regulatory priorities, as Greenberg struggled to answer policy questions during a virtual hearing.
- This AI shortcut could destroy the industry's profits
AI distillation, a technique where models are trained using outputs from other AI systems, is becoming a competitive tool that threatens the profitability of major AI companies. US firms like Anthropic, OpenAI, and Google warn that rivals, including Chinese companies like Alibaba and Z.ai, may use distillation to replicate their models cheaply, undermining investments in data and computing power.
- Anthropic’s Alibaba fight raises a trillion-dollar question for IPO: How defensible is a frontier AI moat against China with Washington’s toolbox?
Anthropic alleges Alibaba used fake accounts to distill its AI capabilities, potentially undermining Anthropic’s competitive edge and IPO prospects. U.S. policymakers are considering updated export controls to address such 'distillation attacks,' with Rep. Michael Lawler’s Remote Access Security Act aiming to restrict foreign access to sensitive AI technology via cloud services.
- Anthropic is accusing China's Alibaba of exploiting its AI models in a large-scale attack
Anthropic accused Alibaba of conducting large-scale distillation attacks to extract its AI model capabilities using 25,000 fraudulent accounts. The company called for stronger legislation against such attacks and cited Alibaba's efforts to replicate US AI advancements without incurring training costs.
- White House talks with Anthropic shift to setting AI security rules
The White House and Anthropic are collaborating on a framework to assess AI security flaws, particularly for Anthropic's Fable 5 and Mythos 5 models, which face export controls due to a perceived jailbreak vulnerability. The effort aims to establish benchmarks for evaluating security risks in AI models, reflecting ongoing negotiations to address disagreements over the severity of the flaw.
- What AI CEOs still don't get about Washington
AI CEOs like Sam Altman and Dario Amodei are pushing for sweeping AI governance policies, but Congress struggles with tech regulation. Their proposals face challenges due to political gridlock and differing priorities between Silicon Valley and Washington.