Surge in AI-Related Cybercrime and Account Exploitation Poses Significant Global Threat

Key Intelligence

• ■The dark web is experiencing a surge in sales of stolen AI account access and Large Language Model (LLM) services, often at heavily discounted rates, indicating a growing black market.
• ■Hackers are employing 'LLM-jacking' techniques to compromise AI accounts and misuse corporate cloud infrastructure for their own operations, including running AI tasks for free.
• ■New sophisticated threats are emerging, such as CARBONATO, the first botnet with an AI agent as its command-and-control engine, operating since October 2024.
• ■Warnings highlight the vulnerability of legacy systems to exploitation by AI agents and the urgent need for innovation in cyber defense to counter these evolving threats.
• ■Researchers are actively identifying and exposing new cybersecurity risks within AI chatbots, underscoring the ongoing challenges in securing AI technologies.