Security Analysis Shows Google Bard Can Help Cybercriminals

A comparative analysis between ChatGPT and Bard Demonstrated some security limitations in Google’s generative artificial intelligence (AI). Check Point Research (CPR) researchers conducted the study and found that the bot helped cybercriminals without encountering any resistance.

CPR, the Threat Intelligence division of Check Point Software, wn managed to generate phishing emails, malicious keylogger and basic ransomware code with Bard. According to the researchers, hackers use this technology to commit crimes, which is something observed in ChatGPT.

“But it appears that Google Bard did not initially learn the lessons from the initial processing of the ChatGPT AI platform on implementing best anti-abuse measures for cyber and phishing abuse of the platform,” the intelligence manager said. The Threat from Check Point Research, Sergey Shykevich.

He also added that the OpenAI robot has received improvements over the past six months, making it harder for cybercriminals to use the robot. Google Bard, however, “is still at a very immature stage from this perspective.” There is an expectation that security measures and limits will be adopted in the near future.

Shykevich ranked the team’s observations by comparing them with ChatGPT. Bard’s anti-exploit restrictors are significantly lower and impose almost no restrictions on the creation of phishing emails. AI develops keyloggers with very little manipulation and also creates the basic ransomware.

“The restrictions that exist at Bard are relatively simple, similar to those observed at the initial launch phase of ChatGPT a few months ago, and still give us hope that these will be stepping stones for a long way and that the platform will embrace the necessary limitations.” and safety limits,” the executive concluded.