Technology news and discussion on TalkFuse.
Anthropic details how criminals and scientists misuse its AI for cyberattacks and surveillance
Anthropic has published a 154-page threat intelligence report detailing how criminals, state-sponsored groups, spyware vendors, scientists and propagandists have attempted to misuse its AI models, including for weapons design, biological research, cyberattacks and surveillance of dissidents. The company said it was disclosing the cases out of a sense of responsibility, describing them not as typical misuse but as the "most notable and novel threat activity" it has identified, and noted it has banned the accounts involved. The report lands two days after Anthropic employee Jacob Coxon resigned publicly, warning that Anthropic and OpenAI were racing towards a "self-improving superintelligence" that could cause human extinction by 2030.
Among the cases were five instances of scientists using Claude for biological research after circumventing safeguards designed to block "unsupported regions" and concealing their intent, including one scientist using the tool to draft a state-sponsored grant application to study the chikungunya virus at what Anthropic identified as a military research institute. Other documented misuse included Russian espionage and "smash-and-grab" hacking operations, a China-based surveillance programme targeting Uyghurs in Syria, propaganda campaigns in Russia, Malaysia, Iran and Bangladesh, and use of Claude in Yemen, China and Russia to help develop firearms, missiles, drones and bombs. Anthropic did not name the institutions or countries involved, citing uncertainty over researchers' intent, and called for the wider AI industry and governments to collaborate on defences as model capabilities grow.
In short
- Anthropic reveals AI misuse cases: weapons design, bioweapons research, surveillance
- Report follows employee's resignation warning of AI "extinction" risk by 2030
- Anthropic banned accounts but withheld names of institutions and countries
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Publishing the cases can help governments, researchers and AI developers understand emerging threats and improve safeguards. Anthropic can reasonably argue that sharing notable examples, while withholding uncertain identities, is part of its responsibility to limit harm and encourage industry-wide cooperation.
The case against
A company’s own threat report is selective and may frame ambiguous research or incidents in ways that serve its interests, so its claims warrant independent scrutiny. Publicising misuse details could also help bad actors adapt, while drawing attention to dramatic cases may distract from the report’s uncertainty and the broader debate over AI risks.
Explained simply
Anthropic is a US-based artificial intelligence company that makes Claude, a tool that can answer questions and help with research and computer tasks. AI systems like Claude can be useful for legitimate work, but their abilities can also be misused, so companies set safeguards to restrict certain requests.
The cases described involve a range of actors, including criminal hackers, state-linked groups, researchers and companies that make spyware. Their reported activities span cyberattacks, surveillance, propaganda and research into weapons or disease; Anthropic says it barred the accounts involved. The company did not identify some institutions or countries, saying it was uncertain about researchers’ intentions.
The disclosures raise questions about how AI companies can detect and limit harmful uses, and what information they should share with governments and other firms. They also sit within wider debate about the risks of increasingly capable AI systems and how quickly they are being developed.
AI Art Culture Cybersecurity Research Science Technology