<ul><li>Researchers have developed a technique to jailbreak multiple AI chatbots, including popular models like ChatGPT and Microsoft Copilot.</li><li>The technique, called 'Immersive World', involves creating a fictional scenario to bypass security controls and develop an effective infostealer malware.</li><li>This highlights the increased risk of cybercriminals with no prior experience in coding being able to create sophisticated malware.</li><li>The rise of AI-powered cyber threats is a serious concern, as it allows criminals to craft more sophisticated attacks with ease.</li></ul>

Not even fairy tales are safe - researchers weaponise bedtime stories to jailbreak AI chatbots and create malware

Discover more