Deepfake generators face diverging objectives between realism and evasion
Vijay Balasubramaniyan · Can We Detect a Deepfake? · Oct 30, 2024 · at 17:33
Vijay Balasubramaniyan, CEO of voice security company Pindrop, explains why adversarial detection testing works against unknown deepfake algorithms.
0:00 / 0:19exact quote · 19.7s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“If a deepfake system has to serve two masters, that is one, I need to make the speech legible and sound as much like Martine, and two, I need to deceive a deepfake detection system. Those two objective functions start diverging.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Vijay Balasubramaniyan
Balasubramaniyan: LLM hallucinations make models ideal tools for fraudsters
“Like in LLMs, hallucination is a problem. So the fact that you're making shit up is a bad idea. But if you have to make shit up to convince someone, LLM is great and perfect.”
Balasubramaniyan: Pindrop detects voice deepfakes with 99% accuracy, 1% false positives
“They're completely detectable. Right now we're detecting them with 99% detection rate with a one percent false positive rate.”
Balasubramaniyan: Audio watermarks get stripped away by telecom transmission
“When you take that audio, play it across air, play it across telephony channels, the bits and bytes, they get stripped away. And so once they get stripped away, an audio is a very sparse channel. So even if you add it over and over again, it's not possible to …”
Balasubramaniyan: AI deepfake detection is 100x cheaper than generation
“It's way cheaper to detect deepfix, right? Because if you think about it, the closest example is Apple released its model, you know, that could run on device. And even that model, which is a small model in order to do lots of things like voice to text and thin…”