AI deepfakes threaten security, experts urges action

Recent incidents show how synthetic media are being used to impersonate top U.S. officials, including Secretary of State Marco Rubio and Trump’s chief of staff Susie Wiles. These fake messages, sent via voice, text, and messaging apps, aim to extract sensitive information or gain access to secure networks.

In one case, a deepfake of Rubio falsely claimed the U.S. would cut Ukraine’s access to Starlink internet services. Another AI-generated robocall mimicked President Joe Biden’s voice to discourage Democratic voters during the New Hampshire primary.

Criminals are also targeting corporate sectors. Using deepfakes, scammers impersonate CEOs or job applicants to access internal systems, steal data, or install ransomware. The U.S. says thousands of North Korean IT workers have infiltrated tech firms using stolen identities—earning salaries and planting malware.

“The financial industry is right in the crosshairs,” said Jennifer Ewbank, former CIA official. “Even trusted colleagues have been fooled.”

Experts believe the solution lies in fighting AI with AI. Tools like Pindrop analyze speech patterns to detect voice cloning. Such systems could soon become standard in interviews or virtual meetings.

Regulators are also calling for stricter laws and improved digital literacy to curb misuse, reports UNB.

“We’ve entered an era where anyone can convincingly impersonate a real person,” said Brian Long, CEO of Adaptive Security.

Despite the challenges, tech leaders remain hopeful. “We’re not going to be subservient to disinformation,” said Pindrop CEO Vijay Balasubramaniyan. “We are going to fight back.”