Hackers are actively looking for ways to turn a chatbot into a cyberweapon.
More and more cybercriminals are looking for ways to use ChatGPT in their activities, according to cybersecurity researchers at NordVPN.
The number of new posts on darknet forums about ChatGPT rose from 120 in January to 870 in February, a 625% increase. The number of new forum threads about ChatGPT rose by 145% - from 37 to 91 in a month - as bot use became the hottest topic on the darknet.
Among the most popular topics on the forums:
how to force ChatGPT to create malware;
how to hack ChatGPT;
how to use it for cyberattacks.
In addition, according to experts from the information security company Norton, the ability of a chatbot to compose texts that are indistinguishable from “human” writing can potentially lead to phishing attacks. On the forums, attackers are looking for ways to force ChatGPT to carry out such attacks by overcoming the content restrictions of the chatbot.
Notably, hackers have already developed a way to bypass ChatGPT restrictions using the chatbot API and are using it to sell services to other cybercriminals so that they can create malware and phishing emails.
NordVPN experts expressed their concern that over the past month, discussions on the dark web have evolved from mere “tricks” and workarounds designed to encourage ChatGPT to do something funny or unexpected, to taking full control of the tool and turning it into a weapon.