The generation of images through artificial intelligence has reached a new milestone with the launch of GPT-4o, which has led to a significant increase in the creation and dissemination of hyper-realistic images on social platforms. Especially referred to as the Studio Ghibli style, these creations have captured the attention of millions, even causing temporary congestion in the AI’s image generation capacity due to their popularity.
Watermarks are coming to AI-generated images
However, the growing popularity of these images raises legal concerns, as their realism could lead to issues regarding copyright and intellectual property. In response to these concerns, OpenAI is considering implementing a watermark on images generated by its AI. This information was revealed by AI researcher Tibor Blaho, who found references in the AI’s code suggesting the inclusion of this watermark, specifically for users utilizing free accounts.
References to image-gen-watermark-for-free indicate that users without a ChatGPT Plus or Pro subscription will see a watermark on their generated images. According to sources consulted by BleepingComputer, those who opt for the premium subscription will not have this restriction, which could encourage more users to pay for the service.
This approach is not unprecedented for OpenAI, which previously raised the possibility of applying watermarks to texts generated by its AI. With the increase in the perfection of artificial intelligence, the implementation of watermarks seeks to mitigate the risks associated with deepfakes and fraud. The current challenge is to find a balance between accessibility in the use of AI and the protection of copyright, while the watermarking system is established as an emerging standard in the field of artificial intelligence.
