In its EWC Submission UN AI Interims Report 03 2024 on the aspect of AI as an “enabler”, the EWC stresses that: “we live in an era in which knowledge is easier to find than ever before. This knowledge is either freely accessible or licensed, and responsible crafted and controlled by human professionals. The portrayal of so-called “AI” as an enabler is a misleading assertion, as it implies that there is a shortage  – but at the same time, it has been observed worldwide that no so-called AI system is reliable, especially in teaching and education. AI means: Fewer quality, fewer trusted works, including educational, cultural and scientific materials.

As “Enablers”, we recommend putting the enforcement of IP into the center:

Regulate Input: As generative systems are based on years of unauthorised, non-remunerated, and non-transparent use of protected works, sanctions are needed, and a remunerated, consent orientated licensing scheme established.

GAI products must be labelled to avoid illegitimate compensation claims, to allocate liability, to avoid unlawfully benefiting from public funding, prizes, scholarships, or project funds.

Authors’ Rights and Copyright legislations must enforce the interests of harmed authors, regarding scraping; storage and reproduction for machine and GAI development. No further exceptions or limitations shall be put in place; consent of authors, and remunerated licensing are the key forward.

Adding the risk of copyright infringement

As regards to AI risks, EWC acknowledges that the Report recognises issues relating to e.g., generative AI and deep fakes, that pose non-revokable risks to societal trust and democratic debate.

However, it is from most importance for the AI Advisory Board to take the following more into account: Generative, analytical and assistive informatics, sub-areas of so-called artificial “intelligence”, threatens numerous jobs and fields of labour in the book sector and will replace some professions by machines in the medium run; be it in the areas of writing, editing, proof reading, production, cover design, illustration, translation, selection and editing of original and translated works, audio book production or in the promotion and distribution of books.

Numerous criminal and damaging “AI business models” have developed in the book sector – with fake authors, fake books and also fake readers. It has been proved that the fundaments for large language models such as GPT, Meta, StableLM, BERT have been illegally built from copyrighted book works and whose sources are often even shadow libraries such as Library Genesis (LibGen), Z-Library (Bok), Sci-Hub and Bibliotik – piracy websites.

Add a Principle Nr 6 and put IP in the center

We currently see a strong underrepresentation of IP, but also personal rights, data protection rights and moral rights, as well as the right of access to cultural creation, in the report.

Since the most successful advanced informatics systems worldwide are currently generative systems that exist on the unauthorised plundering of works, but also of personal data and information of private individuals, and at the same time the technical oligopolies are expanding their power in such a way that access to personal materials must even be granted when using software such as texting, correction or mail programs, image editing or online collaboration tools, in order to use them, we see it as essential that a separate working group is formed on the topic of IP rights. Or, as a minimum standard, IP is a topic in every working group.

Therefore, IP and copyright should be not only an add-on in your precious work, but in the heart of it, and form a Guiding Principle No 6: AI Regulation must put the IP rights of authors at the centre.