Penn State Altoona cybersecurity student Hans Esteban presented his research project on rogue artificial intelligence at the Pennsylvania Capitol on March 19 as part of the Undergraduate Research at the Capitol Pennsylvania Poster Conference. In his research, Esteban’s found that the ChatGPT AI system can create ransomware to infect Windows computers and encrypt user data files.
ALTOONA, Pa. — Penn State Altoona cybersecurity student Hans Esteban presented his research project on rogue artificial intelligence at the Pennsylvania Capitol on March 19 as part of the Undergraduate Research at the Capitol Pennsylvania Poster Conference.
Esteban’s research addressed the possibility that the ChatGPT artificial intelligence (AI) system can create ransomware to infect Windows computers and encrypt user data files. Esteban found that it can.
He used the AI system to code ransomware which effectively encrypted the data files on test computers. In addition, he used ChatGPT to develop a command and control server from which he could remotely manipulate systems that his malware had infected.
Esteban presented to a wide range of people from politicians to the general public, who expressed admiration for his efforts and fear about the potential risks of AI. He noted that their reactions were encouraging, as they showed an increased awareness of generative AI's progression and a readiness to take steps to prevent possible harm.
“The moment when artificial intelligence might turn rogue is uncertain,” Esteban said. “If we fail to act now, we're merely waiting for AI to reach its tipping point, akin to taking a bite of the forbidden fruit.”
Esteban’s faculty advisers for the research project were Syed Rizvi and David Barnes, professors of information sciences and technology.
