Story perspectives
Anthropic blocks malicious Claude uses tied to cyberattacks
By Drooid · · How we work
1 of 2
Story summary
- Anthropic blocked dozens of malicious Claude uses across multiple threat areas.
- It identified Russia-linked group Midnight Blizzard using Claude to steal 300,000 Ukrainian identity records.
- A China-based actor employed Claude to design electronic-warfare and radar-jamming tools against Taiwan.
- Anthropic reported Alibaba routing over 151 million Claude exchanges May-July, peaking at nearly 3 million daily.
- It added stronger safeguards to newer Claude models and shared intelligence with authorities.
1 / 2
