OpenAI has taken a significant step towards ensuring the authenticity of AI-generated images by incorporating the Coalition for Content Provenance and Authenticity (C2PA) specifications into its DALL-E 3 image generation technology. The company announced its plans to add watermarks to images created using DALL-E 3, which will include both visual watermarks and updated metadata, enabling users to verify the image's source via C2PA.
The introduction of watermarks aims to provide users with a quick and reliable method of identifying whether an image has been generated using AI or created by humans. Notably, industry giants like Adobe and camera manufacturers have been adhering to the C2PA standard, emphasizing the importance of authenticity in digital content.
OpenAI has confirmed that the watermarks will be visible on DALL-E 3 generated images across various platforms, including the web, API, and mobile app. The watermark will display details such as the image generation date, accompanied by the C2PA logo positioned at the top left corner of the image.
Currently, access to the DALL-E 3 image generator is exclusive to ChatGPT Plus subscribers, priced at $20 per month.
OpenAI assures users that the addition of watermarks will not impact the performance of AI-generated images in terms of quality or latency. However, there is a slight increase in image size, ranging from three to five percent when generated through the API and up to 32 percent when using the ChatGPT platform.
Despite the presence of watermarks, it is still possible to tamper with the images by physically cropping them or manipulating the metadata. When AI-generated images are captured via screenshots or uploaded to social media platforms, the metadata may be altered or removed.
The initiative to watermark AI-generated images is crucial in addressing various issues associated with such content, including impersonation, deepfakes of celebrities, and the spread of misinformation. By making watermarks easily visible, even to non-tech-savvy audiences, OpenAI aims to enhance transparency and combat the proliferation of misleading digital content.
Furthermore, other tech giants like Microsoft, Samsung, and Meta have also implemented watermarking and metadata updates in their AI-powered image editing tools and platforms, emphasizing the industry-wide commitment to ensuring content authenticity in the digital age.
