On August 2, a new section of the European AI legislation, known as the ‘AI Act,’ will take effect, introducing regulations for general-purpose AI models. The EU has also created an AI code of conduct, which includes non-binding guidelines for implementing these new rules. However, this code must still receive approval from the ‘AI Board,’ comprised of representatives from member states.
Belgium opposed the code, arguing that it lacks sufficient guarantees for copyright protection. Within the EU, copyright is safeguarded not only by the AI Act but also through stronger commitments in the code of conduct, which Belgium believes would better support authors’ rights.
During discussions, Belgium proposed several enhancements, such as simplifying the process for rights holders to opt out, ensuring fair compensation, and prohibiting searches on sites that host illegal content, regardless of commercial intent. Although the revised document shows some improvements, it remains inadequate for Brussels.
“Belgium and other European nations have stressed that this is not the conclusion of the process, and the Commission has recognized that the code may need to be evaluated and amended sooner than currently stipulated,” stated Minister Matz. “I will continue collaborating with the relevant parties to ensure that AI models effectively address the interests of journalists, publishers, producers, and creators.”
As noted on the European Commission’s website, 26 companies, including Google, Microsoft, and Open AI, the creator of Chat GPT, have already endorsed the code of conduct.
Leave a Reply