There are several strategies you can use, but the most important and easiest is that you insert a disclaimer about how your works may be used.
VISDA writes:
Since the summer of 2023, it has been permitted to carry out so-called text and data mining of copyright-protected works on the Internet. The purpose of text and data mining is to find patterns and trends on websites, which can for example be used to improve new technologies. It is debated whether the right to text and data mining also applies to training AI systems.
However, text and data mining of protected content is never permitted if you have made a reservation "appropriately".
Therefore, VISDA recommends that all image creators insert the following caveats on relevant websites:
"Text, graphics, images, sound and other content on this website are protected by copyright law. [insert your company name] reserves all rights to the content, including the right to exploit the content for the purposes of text and data mining, cf. Section 11 b of the Copyright Act and Article 4 of the DSM Directive.”
The reservation should be made in a machine-readable way in metadata and appear in terms and conditions for use of the website.
Please note that the reservation does not apply to research institutions and cultural heritage institutions, which may continue to carry out text and data mining as long as it is done for the purpose of scientific research.
Read more about VISDA's work for image creators at visda.dk: