Google Develops Tool to Identify AI-Generated Images

Google is looking for the best solution to the challenges posed by artificial intelligence (AI). As we all know, AI-generated images have become commonplace on the internet, not just social media. Media companies have started to regulate the use of such content to avoid potential copyright infringements. But it’s safe to say that some AI technologies can make it difficult to tell whether a photo is AI-generated or manually developed from scratch.

Google wants to make this process easier for people and is using its AI division, DeepMind, to create a tool that can detect AI-generated images and even add a watermark to them to help people instantly identify the image’s origin.

The AI division has released a beta version of software called Synth ID, which adds a watermark to pixels in an image that may not be visible to our eyes, but the technology can identify the image’s identity.

Since the software is in beta, DeepMind can continue to test new unreleased features with a limited number of users and companies. Google uses its very ingenious deep learning models to design the watermarking tool, and the company claims that adding a mark has no effect on image quality, color levels, or other aspects.

While Synth ID’s efforts are a step in the right direction, the pace of AI technology development, the industry needs more improved solutions to help avoid major AI misinformation disasters, especially in cases involving Deepfakes. With major government elections coming up in the next 12 months, it’s imperative that Synth ID’s beta model delivers on its promise and helps the world avoid the dangers of AI malpractice.

Google has talked about being in charge of regulating AI, but in its ongoing efforts to dominate the segment, other players are needed to make these products and tools easier and faster to use.