Stolen AI Accounts Traded at Bargain Prices on Dark Web as 'LLM Jacking' Spreads

Hackers Steal AI Login Credentials
Resell for Up to 97% Less Than Regular Price
Offer "Connection Guarantee" Service Against Account Blocks

It has been revealed that there has been a sharp increase in new types of cybercrime where hackers hijack companies' artificial intelligence (AI) accounts and servers and sell them at bargain prices on the dark web.

On the 27th (local time), the Financial Times (FT) reported, citing John Hultquist, chief analyst of Google's Threat Analysis Group, that since the beginning of this year, there has been a significant surge in cases of buying and selling AI model login credentials on the dark web, as well as cases where hackers occupy companies' computational resources. Chief analyst Hultquist referred to this as "large language model (LLM) jacking," explaining, "A new economy is forming in the cybercrime underground market, centered on access rights to AI."

According to reports, dark web marketplaces are trading access to AI models from companies such as Anthropic, Google, and OpenAI at discounts of up to 97% off the normal price. While subscription fees for premium AI models like ChatGPT and Claude can reach up to $200 (270,000 won) per month, FT noted that on these black markets, stolen accounts are available for just a few dollars.

Some sellers even offer a "connection guarantee" service, promising to provide new login credentials at no extra cost if a stolen account is blocked. Because AI companies monitor for signs of account abuse and cut off access when issues are detected, these cybercriminals actively market new accounts that circumvent such blocks.

FT reported that hackers who acquire high-performance AI at significantly reduced prices are also minimizing the costs involved in carrying out cyberattacks. Companies defending against such attacks have to pay regular high prices for AI usage, while hackers can leverage the same technology much more cheaply. Analyst Hultquist pointed out, "Attackers gain a cost and efficiency advantage by being able to acquire AI tokens at a much lower price."

There are growing calls for companies to not only expand the adoption of AI, but also to quickly overhaul their related security systems. Hultquist warned, "Those who consider AI as just a temporary trend and neglect proper safeguarding may one day find themselves in deep trouble," adding, "There is a high risk of facing security incidents, warnings, and attacks at an unprecedented scale." He emphasized, "Internal security systems must be updated immediately."

This content was produced with the assistance of AI translation services.