The moment we found out that ChatGPT could produce code, we all knew that hackers were going to use it. As such, the number of malicious software on the internet has exploded. However, there is some good news here. Hackers have been using Google Gemini for their own nefarious needs, but they’ve been failing.
As you can tell, Google has some pretty tight safeguards on Gemini. Its models are some of the most powerful on the market, so you know that someone can do some major damage if they use it without any guardrails. All of the major AI models have some sort of guardrails to make sure that people aren’t causing havoc.
Hackers have been using Google Gemini but failed at getting what they want
Google doesn’t exactly spy on people who use Gemini. However, if you’re trying to use Gemini to teach you how to hack a corporation or build a bomb, you might be on Google’s naughty list. The company reported that it found several hacking groups using Gemini to help them with their hacking skills.
10 groups are located in Iran, 9 in North Korea, 20 in China, and some in Russia. All of these groups tried using Gemini for their dirty deeds, but thanks to the safety protocols that Google put into place, they couldn’t achieve more than some “productivity gains“.
“Some malicious actors unsuccessfully attempted to prompt Gemini for guidance on abusing Google products, such as advanced phishing techniques for Gmail, assistance coding a Chrome infostealer, and methods to bypass Google’s account creation verification methods,” Google stated.
So, it seems that they tried to build certain malicious tools and use Gemini, but they couldn’t. Rather than using Gemini to help them hack, they mostly used it for research. Google stated that it was hit the hardest by Iranian threat actors, and they used Gemini for “research on defense organizations, vulnerability research, and creating content for campaigns.” Far from generating malware, they used it as a research tool.
Other groups would use Gemini for translating content, localizing content, fixing code, and more. It’s the equivalent of buying a robot in the hopes of it cleaning your house but finding out that it can only help you sweep.
This shows how good Google is at keeping its AI tools safe from hackers. This is extremely important in this day and age, as AI tools are becoming more accessible.
