Anthropic accused Alibaba of mass-scale model distillation, while both Anthropic and OpenAI faced US government pressure over model releases, reshaping the competitive landscape around security and geopolitical trust.
After weeks of negotiations, the White House authorized Anthropic to restore access to its most advanced model, Mythos, for a select group of US companies and government agencies. A decision that highlights the increasingly strategic role of frontier...
The Trump administration has asked OpenAI to release its next LLM, internally called GPT-5.6, only to a short list of trusted partners with customer-by-customer approval. The move marks a turning point in the AI control debate, with direct implicatio...
A campaign with 28.8 million exchanges and nearly 25,000 fraudulent accounts targeted Claude's most advanced capabilities. The incident highlights the vulnerability of API-exposed models and strengthens the argument for on-premise deployments in sens...
Anthropic has publicly accused Alibaba of a ‘brazen’ campaign to illicitly extract AI capabilities, likely via model distillation. The case, covered by CNBC and Bloomberg, raises urgent questions about cloud API security and IP protection. AI-RADAR e...
Anthropic claims Alibaba carried out a large-scale distillation attack on Claude, escalating the US-China AI conflict. The incident raises critical questions about LLM security and intellectual property protection, with direct implications for organi...
The White House reportedly asked to replace Anthropic CEO Dario Amodei, called a "weirdo" by an official, with co-founder Tom Brown in high-stakes AI policy meetings. A sign that political relationships now weigh as much as the technology itself.
An AP source says Anthropic’s top model, Mythos, surfaced vulnerabilities in classified US networks within hours. The exercise revives the debate over on-premise AI, data sovereignty, and the true cost of deployment in sensitive environments.
