Home · Technology · Sep 29 archive
Anthropic Joins ARPA-H AI Initiative, Warns of Chinese Distillation Attacks
Confirmed
In Short: In a related development, Anthropic has released a report detailing persistent distillation attacks by Chinese AI companies, including Alibaba. Distillation attacks involve copying the behavior of AI models to create cheaper, less secure versions.
The company will hold a closed-door event focused on healthcare applications of AI, according to Decrypt. This move follows a significant market impact in January when a Chinese startup launched a competitively priced AI model, reducing Nvidia's market capitalization by approximately $600 billion.
Anthropic CEO Dario Amodei stated that the company will provide outside evaluators with permanent, staff-level access to ensure the safety of its AI systems. This transparency measure is aimed at addressing concerns about the risks posed by AI technology.
In a related development, Anthropic has released a report detailing persistent distillation attacks by Chinese AI companies, including Alibaba. Distillation attacks involve copying the behavior of AI models to create cheaper, less secure versions.
The report, released Thursday, highlights that these attacks have intensified as competition in the AI sector has grown. Anthropic describes the latest campaigns as the largest and most aggressive it has observed.
France organized a session at the United Nations Security Council, where leaders from Anthropic, OpenAI, and other AI firms briefed the 15-member council on AI risks. The council has the authority to authorize sanctions and military action.
The UN session included presentations from OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, AI safety researcher Yoshua Bengio, and Hugging Face CEO Clément Delangue, emphasizing the global nature of AI safety concerns.
Additionally, Anthropic will participate in the Disrupt 2026 conference, joining other leading AI companies such as OpenAI and Replit. The event will feature discussions on the latest advancements in AI technology.
What this adds
The sources have not established the exact financial impact of the distillation attacks on Anthropic or other companies.
What's confirmed
- The company will hold a closed-door event focused on healthcare applications of AI, according to Decrypt. This move follows a significant market impact in January when a Chinese startup launched a competitively priced AI model, reducing Nvidia's market capitalization by approximately $600 billion.
- Anthropic CEO Dario Amodei stated that the company will provide outside evaluators with permanent, staff-level access to ensure the safety of its AI systems. This transparency measure is aimed at addressing concerns about the risks posed by AI technology.
- In a related development, Anthropic has released a report detailing persistent distillation attacks by Chinese AI companies, including Alibaba. Distillation attacks involve copying the behavior of AI models to create cheaper, less secure versions.
- The report, released Thursday, highlights that these attacks have intensified as competition in the AI sector has grown. Anthropic describes the latest campaigns as the largest and most aggressive it has observed.
- France organized a session at the United Nations Security Council, where leaders from Anthropic, OpenAI, and other AI firms briefed the 15-member council on AI risks. The council has the authority to authorize sanctions and military action.
- The UN session included presentations from OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, AI safety researcher Yoshua Bengio, and Hugging Face CEO Clément Delangue, emphasizing the global nature of AI safety concerns.
- Additionally, Anthropic will participate in the Disrupt 2026 conference, joining other leading AI companies such as OpenAI and Replit. The event will feature discussions on the latest advancements in AI technology.
What's still developing
- The model is open-weight, priced at roughly a third to a fifth of its US frontier rivals, and reportedly cost a fraction of what OpenAI or Anthropic spend to train comparable systems.
- The Chinese startup that wiped roughly $600 billion off Nvidia's market cap when it launched a cut-price model in January 2025 has been asked to make a statement, alongside another Chinese firm, Moonshot, developers of the Kimi models..
- France organized Wednesday's session and holds the council's rotating presidency this month.
- Anthropic previously spoke out about distillation attacks in February, even calling out specific labs.
- The bulk of the distillation attempts came from a campaign attributed to Alibaba, which Anthropic describes as the largest wholesale distillation effort the company has ever observed.
- Anthropic typically does not make its models’ internal chain of thought available to users, instead displaying “summarized thinking” blocks that give a general overview.