Dark Web Sells AI Model Access at 97% Off — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Google Threat Intelligence Group identified dark web marketplaces reselling access to AI models at discounts of up to 97%, including services from Anthropic, Google, and OpenAI.
- Security researchers warn of a surge in 'LLM-jacking' attacks, a term for criminals hijacking companies' costly AI resources to sell on underground markets.
- The stolen access spans models from three major frontier AI labs, underscoring how expensive compute and API pricing have made legitimate AI services a target for resale.
Why it matters: Frontier AI labs now face a monetization layer for stolen access: criminals are undercutting legitimate API pricing by up to 97%, potentially driving up compute costs and abuse risk for Anthropic, Google, and OpenAI while their infrastructure picks up the tab.
Ask SkimNews
