In a recent development that has significant implications for cybersecurity, researchers have successfully designed a “worm” that is driven by Artificial Intelligence (AI). This advanced form of malware showcases the potential for AI to be utilized in creating more sophisticated cyber threats.
According to a report by WIRED, the research team consisting of Ben Nassi, Ron Bitton, and Stav Cohen conducted a series of experiments to demonstrate the capabilities of their AI-powered malware. Their findings revealed an alarming success rate on attacks to email systems powered by Gemini, OpenAI, and others, raising concerns that it could potentially get personal and sensitive information like names, numbers, addresses, and more.
The worm, dubbed as Morris II, works by “using a text-based self-replicating prompt and by embedding a self-replicating prompt within an image file.”
See also: How Android malware slips inside the Google Play Store
The use of generative artificial intelligence in developing malware represents a paradigm shift in how cyber threats are created and deployed. Traditional malware relies on predefined scripts and can often be mitigated by keeping software up-to-date and using reliable antivirus solutions. However, the introduction of AI into this domain introduces an adaptive and learning component, making the malware far more elusive and challenging to defend against.
While the research is primarily academic, it underscores the urgent need for the cybersecurity industry to quickly adapt its defense mechanisms and prevent these kind of attacks.