Stealing Your Voice: How Generative AI Audio Powers Scams, and What You Can Do About It
All information provided below is strictly hypothetical. It is presented for educational purposes only. The goal of this blog post is to help individuals and businesses defend themselves from this new attack vector. AI will only become more prevalent. Learning how to defend oneself will only become increasingly important.
The rise of generative AI has resulted in incredible advancements in the generation of both audio and video. This has enabled meaningful improvements in text to speech applications, translation, and the generation of video content for creators on YouTube or other content focused sites. But with this comes a dark side, as the ability to generate realistic audio and images, comes with the opportunity to more effectively mimic the voice and appearance of the ones you love.
This is already happening. Celebrities across the internet are finding that their “voice” has been used to endorse products they’ve never heard of, and music has been created in their name that they never took part in. Joe Rogan as an example, has found himself the unwilling endorser of a “libido booster” as well as Drake who was surprised to find out he had released a new song with The Weeknd. While these examples might seem small, or even a bit catchy (sorry Drake), they’re proof that it’s possible to realistically spoof the voice of another human being to make them express themselves in a way they might not support. Attackers are using their celebrity to gain clout for products or music that otherwise would have less attention around them. In short, malicious marketing.
