A low-tech solution from the past may be your best defense against AI deepfakes (www.zdnet.com)

Recent incidents involving advanced AI-generated deepfakes have raised significant concerns about cybersecurity, particularly following a highly sophisticated scam that cost the professional services firm Arup $25 million in fraudulent wire transfers. An AI clone of the company's CFO was used to trick an employee into authorizing these transactions, illustrating that traditional verification methods like video calls and voice recognition are now unreliable. Experts, including Deepak Gupta from GrackerAI, emphasize that the reliance on visual and auditory cues for identity verification is increasingly compromised, making it clear that organizations need to adapt their security protocols. To combat this evolving threat, cybersecurity professionals recommend a return to low-tech solutions, such as implementing verbal passphrases for authenticating calls. These passphrases, when used consistently across the organization, serve as an effective barrier against deepfake attacks that exploit traditional security measures. Additional strategies suggested include customizing passphrases for various roles, rotating them unpredictably, and enhancing them with other authentication methods like out-of-band callbacks. By refining security systems to target high-risk interactions rather than applying broad measures, companies can foster a culture of vigilance, thereby enhancing their defenses against sophisticated cyber threats.