According to recent reports, an OpenAI DALL-E 3 image detection tool is in the works. This tool will help spot or identify AI-generated images from the DALL-E 3 system. Currently, the AI tool DALL-E 3 is available to millions of users and creates impressive images using prompts from users.
These images are then put to use by the generator for various reasons, and some bad actors are putting this system to negative use. With the US presidential election drawing closer, some might intend to use this AI tool to share misleading details with millions of unsuspecting people. For this reason, the artificial intelligence firm is working on a system to identify AI-generated images from DALL-E 3.
According to OpenAI, this tool has a 98% accuracy when it comes to the identification of DALL-E 3 AI-generated images across the web. Even if some modify these images to trick the system, it can still find its way around the modifications and identify the image. This tool, once available, will be in use across a bunch of online platforms where misinformation can spread to keep users of these platforms safe at all times.
To further protect users, OpenAI plans on adding tamper-resistant watermarks to DALL-E 3 images. If anyone generates an image using a prompt on DALL-E 3, the system will automatically add a watermark pointing out that the image is AI-generated. This will help those who come in contact with the image while surfing the internet to understand that it isn’t a real image regardless of how convincing it looks.
Aside from its use in curbing misinformation during the US presidential election, this OpenAI DALL-E 3 image detection tool will find good use in other areas. With its application, spotting AI-generated images across the media will be easy, hence helping to clear any misinformation. Once this tool is available, OpenAI might make it available for use via their platform for easy access to users.