The fascinating world of hackers and unbridled artificial intelligence on the Dark Web
Hackers are increasingly exploiting generative artificial intelligence (AI) to carry out their criminal activities. A Kaspersky Dark Web investigation found that the use of AI, particularly generative AI tools, has become common and concerning.
Thousands of discussions on the use of AI for illegal purposes
Kaspersky Digital Footprint Intelligence, a Russian cybersecurity company, analyzed the Dark Web to identify discussions about the use of AI by hackers. Researchers found thousands of conversations discussing the use of AI for illegal and malicious purposes.
During the year 2023, no less than 3,000 discussions were recorded, with a peak in March. Although these discussions tend to diminish over the year, they remain present and active on the Dark Web.
AI at the service of cybercriminals
These discussions mainly revolve around malware development and illegal use of language models. Hackers are exploring avenues such as processing stolen data, analyzing files from infected devices, and many others.
These exchanges demonstrate the growing interest of hackers in AI and their desire to exploit its technical possibilities in order to carry out criminal activities more effectively.
Selling Stolen ChatGPT Accounts and Jailbreaks on the Dark Web
Besides discussions about the use of AI, the Dark Web is also a thriving market for the sale of stolen ChatGPT accounts. Kaspersky has identified more than 3,000 ads selling paid ChatGPT accounts.
Hackers also offer automatic registration services to massively create accounts on demand. These services are distributed over secure channels like Telegram.
Additionally, researchers have seen an increase in the sale of jailbroken chatbots such as WormGPT, FraudGPT, XXXGPT, WolfGPT, and EvilGPT. These malicious versions of ChatGPT are free from limitations, uncensored and loaded with additional features.
A growing threat to cybersecurity
The use of AI by hackers represents a growing threat to cybersecurity. Language models can be exploited maliciously, increasing the potential number of cyberattacks.
It is therefore essential to strengthen cybersecurity measures to counter these new forms of AI-based attacks. Experts must remain vigilant in the face of these constant developments and work to develop effective strategies to counter cybercriminals.