Striking Move from OpenAI: Artificial Intelligence Generated Images Can Now Be Detected!

Striking Move from OpenAI: Artificial Intelligence Generated Images Can Now Be Detected!

OpenAI announced that it is working on a new tool that can detect images created with DALL-E. Here are the details!

OpenAI is working on a new tool to detect whether an image was made with DALL-E 3 and a new watermarking method to more clearly mark the content it produces.

In a blog post on the subject, OpenAI announced that it has started to develop new methods to examine content and prove whether it was made by artificial intelligence. These include a new image detection classifier that uses artificial intelligence to determine whether a photo was made by AI.

The new tool basically estimates the likelihood that an image was made by DALL-E 3. OpenAI claims that the tool works even if the image has been cropped, compressed, or had its colors altered.

The Tool Cannot Yet Detect Midjourney Images

While the tool can identify whether an image was made with DALL-E 3 with about 98% accuracy, its performance in identifying whether an image is from other AI models is not as good, identifying only 5 to 10% of images from other image-generating AIs such as Midjourney.

The company has also begun adding watermarks to Voice Engine text-to-speech clips, which are still in the testing phase, to make it clear that AI-generated content is AI-generated.

OpenAI has been working on detecting AI-generated content for years, and the company’s work in this area will continue.

So what do you think about this issue? You can share your thoughts with us in the comments section below.