Yes, such a tool exists, and I am delighted to see that my images have indeed been used. I would have been disappointed otherwise. Here’s why.

I support the principle of Open Source and accept that my images available on the Internet can be used for AI training. This means I contribute to the creation of powerful tools that surpass our individual capabilities and serve humanity. These tools require vast amounts of data to best represent the human mind and our society. However, I refuse to let my style be freely exploitable by other artists for generating images. This distinction is crucial to protect artistic integrity.

It is important to distinguish between two phases in the AI process: training and inference (the generation of responses). We have surpassed the first training phase, which required a lot of general data from the web. Now, AI must train on high-quality human data to achieve unprecedented excellence in digital content. The recent agreement between OpenAI and Le Monde is a clear example of this.

During training, my photos represent only a tiny fraction among billions of images, making exact reproduction improbable, except in very specific cases I have already mentioned in a previous post.

The inference phase is more complex and requires deep ethical reflection. Unlike humans, AI has no subjectivity or aesthetic desires. It generates works through an algorithmic process, without intrinsic artistic intention. The copyright issue arises when AI is explicitly guided to imitate a particular artist's style.

The HaveIbeenTrained tool, which I presented several months ago at Mazarine, allows artists to check which of their works have been used for AI training and to exercise their right to opt out of future trainings.

Moreover, this quest for quality data could fund fully human research and promote collaborations between ethicists, artists, anthropologists, and others to create human training content, thus maintaining an active role in the orientation of AI. Art remains above all a means of expression and advocacy, concepts foreign to AI.

Ultimately, the integration of AI in the creative field must be guided by deep ethical reflection, ensuring that our technological progress enriches our humanity without diluting it.

Dimitri Daniloff

Link https://haveibeentrained.com/