עובדים מתחזים מבוססי בינה מלאכותית מנסים לחדור לחברות הייטק

As if deepfake videos weren't already enough to worry about, now there are fake job applicants using AI-generated videos to get through remote interviews. When hired, these fake people can do incredible damage to a company, like planting malware and demanding ransoms. NTT Sean Marshall has more for us. Yes, yes, it's much better. Thank you for joining again. Imagine hiring an AI-generated fake person who appears to be the perfect fit for a role at your company. Being tricked through a video conference call sounds like something out of a science fiction movie, but this is exactly what information security company PinDrop has been finding. They offer services to detect AI fraud. PinDrop describes a recent example, a job application The candidate they call Ivan X appeared well-qualified on paper, but the company says there were several red flags during the video interview. His facial expressions seemed slightly out of sync with his words, a telltale sign of deepfake video manipulation. His voice occasionally dropped out or did not align perfectly with his lip movements. And when the interviewer asked an unexpected technical question, the candidate paused unnaturally, as if processing the response before playback. PinDrop's AI fraud detection technology confirmed that they were face-to-face with a deepfake candidate. These deepfake personas will go to great lengths to apply at companies, which includes making fake LinkedIn profiles, AI-generated resumes and cover letters, and fabricated work history and references. Once inside a company, deepfake employees can hold internal systems hostage, locking critical files and demanding ransom payments. They can also commit payroll fraud and steal intellectual property. But PinDrop says that wasn't the last of Ivan X. The fake applicant resurfaced again eight days later after they flagged the original as fraudulent, but they let the interview go ahead anyway. The startup used a new video authentication program it created to confirm he was a deepfake fraud. Ivan claimed to be located in western Ukraine. His IP address indicated he was actually thousands of miles to the east, possibly in a Russian military facility near the North Korean border. Deepfake people are no longer science fiction. The CEO of a company that goes after crypto criminals told CNBC that hundreds of North Korean spies regularly apply for sensitive positions at her firm. Research company Gartner predicts that by 2028, one in four candidate profiles worldwide will be fake. Sean Marshall, NTD News.