ChatGPT is now a tool that hackers and threat actors utilize to develop malware and other accessing programs used to steal from people, using OpenAI's API via a GPT-3 model. This underground version of the experience centers on bots that focus on accessing these OpenAI models to infiltrate the service using Telegram to improve their codes for stealing.
There are countless applications of the famed chatbot from OpenAI, but this is a next-level development for the highly assistive tech that was once used for creating content.
ChatGPT: Hackers Use OpenAI's System to Create Malware
(Photo : NICOLAS MAETERLINCK/BELGA MAG/AFP via Getty Images)
CheckPoint Research announced via a blog post that cybercriminals now use a tactic that bypasses OpenAI's API on its GPT-3 creation, one that helps improve their code or create malware to steal from people. More specifically, this is the "text-davinci-003," similar to ChatGPT, which runs on the same GPT-3 model.
The creations are done via an underground channel in Telegram, and hackers found a way to turn this to their advantage to create codes to write malware, phishing tech for e-mails, and more.
Apart from asking the tech to create codes from scratch, threat actors use this bad API opening to "improve" old codes and instigate their hacking campaigns.
The researchers' previous blog demonstrated how it used ChatGPT to create a full infection flow via a spear-phishing e-mail to deploy reverse shells, which only need English commands.
Read Also: ChatGPT Answers if It Will Cause Mass Unemployment: 'Unlikely'
OpenAI API GPT-3 Access via Telegram Bot
OpenAI's integration of its GPT-3 API and ChatGPT on Telegram Channels via bots have paid its price, as it has limited restrictions that threat actors could manipulate and bypass. The illicit activities are also sold here, being a malware service that people may use to execute against their targets.
ChatGPT is a Tool Used to Cheat
There are many speculations that OpenAI's famed text content generator may soon be used to create more than essays or scripts, as it could also be used for illicit behaviors. One of the main concerns now is using ChatGPT for schoolwork, with many schools already banning the tech for students to use in creating their homework or projects.
Because of ChatGPT's diversity and immense knowledge of different topics, some students' primary focus is to do their work for them. However, professors warn that its use would not result in good grades immediately, and despite its outstanding content generation, it is still subject to checking and review.
Before ChatGPT was used to create malicious codes to help develop malware, its application was already made to create dishonest work against educators and institutions to get better grades. However, the fears of its creators for the tech to be used for hacking people is now real, and the only way to prevent this is to improve their restrictions against bypassing their API.
Related Article: Bill Gates Says AI, ChatGPT Progress 'Excites' Him
