The surge in popularity of voice messages, particularly among the younger demographic, has attracted the attention of malicious actors seeking to exploit this trend for their own benefit. AI-generated audio deepfakes have emerged as a significant threat in the realm of cyberattacks, particularly those involving identity theft. Account takeover attacks, in which hackers seize control of accounts to gain access to valuable information, are particularly prized by cybercriminals. Aaron Painter, CEO of security solutions provider Nametag, explains the allure of such attacks: “Taking over someone’s account gives you control over everything that account has access to. For an employee account, this could mean planting ransomware or accessing sensitive company data. For a customer account, the hacker could hijack social media or bank accounts. Deepfakes provide bad actors with unprecedented capabilities.”

The proliferation of audio deepfakes can be attributed to the ease with which they can now be created. While in the past, it may have taken 20 minutes of audio to produce a convincing deepfake, today, only a few seconds are required. Painter notes that while higher-quality deepfakes may necessitate more audio data, even a low-quality deepfake can be effective in a cyberattack.

Jason Glassberg from Casaba Security anticipates that the next wave of cyberattacks will exploit the habits of the younger generation, who frequently communicate through voice notes. While individuals may exhibit caution in text-based communications regarding phishing attempts, they may be more susceptible to deception in voice messages. Glassberg highlights the persuasive nature of voice messages, noting that they can be more convincing than written communication.

In determining whether an audio message is a deepfake, Glassberg proposes evaluating the context of the message, considering the identity of the sender, and the communication channel being used. Messages in public group chats may raise more suspicion than private, one-on-one exchanges. He also advises listeners to be mindful of editing artifacts, unnatural noises, out-of-character statements by the speaker, and the absence of breathing, which is often absent in deepfake voices.

Overall, the prevalence of AI-generated audio deepfakes poses a significant cybersecurity risk, particularly as malicious actors continue to exploit emerging communication trends. As technology advances, it is crucial for individuals and organizations to remain vigilant and adopt measures to detect and mitigate the threat of audio deepfake attacks.

Извор линк