Transparency, cultural diversity, equal pay, copyright protection were all issues discussed at the round table entitled “Il ruolo dell’intelligenza artificiale nell’industria audiovisiva: opportunità, sfide ed implicazioni per il diritto d’autore e la tutela di creatori e artisti” (The role of AI in the audio-visual industry: opportunity, challenges and implications on the copyright and protection of creators and artists) organised by DGCA – MiC, in collaboration with the European Commission’s Directorate?General for Communications Networks, Content and Technology, which gathered together institutions and key legal and industry figures in the setting of the 81st Venice Film Festival.
The AI issue is one that has been discussed several times in the past days by the industry figures present at the Lido because its rapid progression has begun to significantly change the landscape of the audio-visual industry. From screenwriting to post-production, AI technology – especially generative tools – is creating both opportunities and challenges for the market.
From the copyright perspective, the emergence of generative AI tools raises two issues: the use of content protected by copyright as data to “train” AI modelling (“input”) and the potential application of copyright to the output generated by (or with the assistance of) AI technology (“output”).
With regards to input, the 2019 directive on copyright in the single digital market introduced two exceptions for the approach to text and data mining (TDM). In particular, article 4(3) of the directive provides an important balancing measurement to allow rights holders to keep control of their content (e.g., to negotiate their commercial use) or to exclude them completely from TDM processes, including those underlying AI generative tools (opt-out options).
The AI Act includes measures that aim to improve transparency around AI training data and to support respect of EU copyright norms. It introduces two obligations for the providers of AI models for general purposes (GPAI):
a) To respect copyright in the EU, in particular to identify and respect rights that opt out of TDM processes;
b) To elaborate and make available to the public sufficiently detailed summaries about the content used for training, using the model supplied by the European AI Office.
With regards to output (the generation of content by AI), European copyright law links the protection of copyright to the result of human effort that demonstrates the “intellectual creation of the writer”. Therefore, if GenAI tools create output without human creative choices, that output is excluded from copyright protection. However, the issue of whether and to what extent content generated with AI assistance should be afforded copyright protection is still under discussion.
In Italy, a draft law, approved in April, proposes to realign national legislation with responsible use of AI. The draft aims to integrate the European norm, requiring that the use of AI systems in creating text, photographic, audio-visual and radio content be identified and notified.
At the same time, there is also the proposal for copyright to protect works generated with the assistance of AI. The text aims to update present domestic copyright policy, clarifying that the protected “works of invention” must be “human” in origin and that even works created with the use of AI tools may be protected by copyright as long as their creation derives from the writer’s intellectual work.
The Copyright Committee of the Ministry of Culture, after laying out a position document at the start of the year on the impact of AI on copyright, has been called upon to provide consulting support for the proposed law, underlining the need for balancing requirements for expansion in the digital economy with protection and incentives for human creativity. The Ministry’s new Tax Credit law includes norms relating to the eligibility of above the line costs linking to the use of AI and transparency requirements.