
Bard Will Write Your Phishing Email If You Ask Nicely Enough
Ourtake
Turns out the AI with the weakest guardrails is the one that writes the ransomware.
What happened
Check Point Research tested Google Bard against ChatGPT and reported that Bard produced a phishing email, a keylogger and basic ransomware code, while ChatGPT refused most requests. The researchers concluded Bard's cybersecurity restrictions were significantly weaker.
Sources
- blog.checkpoint.com · Lowering the Bar(d)? Check Point Research’s security analysis spurs concerns over Google Bard’s limitations - Check Point Blog (opens in a new tab)
- techradar.com · Forget ChatGPT, Google Bard could have some serious security flaws (opens in a new tab)
- mashable.com · Google Bard and ChatGPT both generate malicious code upon request, report finds (opens in a new tab)