This is how the world is Torn apart by AI-generated deepfakes, and recent half-hearted attempts to stop them are of no avail.federal regulator Deepfake robocall made illegal on thursday, like them Impersonating President Biden in New Hampshire primary. meanwhile, OpenAI and Google publish watermark This week we’ll be adding AI-generated labels to images. However, these measures lack the power needed to stop AI deepfakes.
“They are here to stay,” said Pindrop CEO Vijay Balasubramanyan, who identified Eleven Labs as the service used to create the fake Biden robocalls. He said, “Deepfake detection technology needs to be adopted at the source, the transmission point, and the destination. It just needs to happen across the board.”
Efforts to prevent deepfakes only scratch the surface
According to Balasubramaniyan, the Federal Communications Commission’s (FCC) outlawing of deepfake robocalls is a step in the right direction, but there is minimal clarity on how this will be enforced.Currently we are Catch deepfakes after they are compromised, and rarely punishes the villains responsible. It’s too slow and doesn’t actually address the problem at hand.
OpenAI this week introduced watermarks embedded in visual and photo metadata to Dall-E images. However, at the same time, the company I admitted that this can be easily avoided by taking a screenshot. This felt less like a solution and more like the company saying, “Well, at least we tried!”
Meanwhile, a deepfake of the boss of a Hong Kong financial official I scammed him out of $25 million.. This is a shocking incident that shows how deepfake technology is blurring the lines of reality.
The deepfake problem will only get worse
These solutions alone are not enough. The problem is that deepfake detection technology is new and hasn’t caught on as quickly as generative AI. Platforms like Meta, X, and even telcos should embrace deepfake detection. These companies are making headlines for all their new AI capabilities, but what about their AI detection capabilities?
If you’re watching a deepfake video on Facebook, you should see a warning about it. If you receive a deepfake call, your service provider must have the software to catch it. These companies can’t just fold their hands, but they’re certainly trying.
Deepfake detection technology also needs to further improve and become more widespread. According to CopyLeaks CEO Alon Yamin, deepfake detection is currently not 100% accurate at anything. His company has one of the best tools for detecting AI-generated text, but detecting AI audio and video is an entirely different challenge. Deepfake detection lags behind generative AI and needs to be strengthened quickly.
deepfakes are real new misinformation, but it’s much more convincing. While there is hope that technology and regulators are catching up to address the problem, experts agree that deepfakes will only get worse before they get better.
