From Hacks to Bioweapons, Claude Misuse Is Now Everywhere
wired ·
In its final weekly security roundup, WIRED reports that Anthropic released an eight-month report documenting widespread misuse of Claude. The report described state-sponsored and cybercriminal hacking, disinformation campaigns, and attempts to develop bioweapons such as pathogens and toxins, with Anthropic saying it disrupted the activity. It specifically cited Russian state-sponsored hackers known as Midnight Blizzard using Claude for reconnaissance and breaching Ukrainian and other European government networks, and ShinyHunters using it across hacking and extortion campaigns. Separately, the US government seized Telegram channels and sanctioned Xinbi Guarantee, described as the internet's largest illicit marketplace, with over $30 billion in sales mostly tied to money laundering for pig-butchering crypto scams, and announced raids on 13 scam compounds in Madagascar. A Ukrainian member of the Conti ransomware gang, Oleksii Oleksiyovych Lytvynenko, was sentenced to four years in prison. Research found Meta failed to catch roughly 350 AI-generated child abuse ads, some showing real children including a European royal family member, prompting legislative scrutiny and an order from the San Francisco City Attorney's Office. The roundup also noted Meta's new personal AI agent and a proposed class action over photo harvesting, Clearview AI's unreported InquiryIQ prototype, Apple's new audio intelligence features, a US-Mexico laser drone operation, and a GTA V mod for destroying in-game Flock cameras.
wired