LONDON — Representatives of the creative industries are urging legislators not to water down forthcoming regulations governing the use of artificial intelligence, including laws around the use of copyrighted music, amid fierce lobbying from big tech companies.     

On Wednesday (Dec. 6), policy makers from the European Union Parliament, Council and European Commission will meet in Brussels to negotiate the final text of the EU’s Artificial Intelligence Act – the world’s first comprehensive set of laws regulating the use of AI.  

The current version of the AI Act, which was provisionally approved by Members of European Parliament (MEPs) in a vote in June, contains several measures that will help determine what tech companies can and cannot do with copyright protected music works. Among them is the legal requirement that companies using generative AI models like OpenAI’s ChatGPT or Anthropic’s Claude 2 (classified by the EU as “general purpose AI systems”) provide summaries of any copyrighted works, including music, that they use to train their systems.

The draft legislation will also force developers to clearly identify content that is created by AI, as opposed to human works. In addition, tech companies will have to ensure that their systems are designed in such a way that prevents them from generating illegal content.

While these transparency provisions have been openly welcomed by music executives, behind the scenes technology companies have been actively lobbying policymakers to try and weaken the regulations, arguing that such obligations could put European AI developers at a competitive advantage.  

“We believe this additional legal complexity is out of place in the AI Act, which is primarily focused on health, safety, and fundamental rights,” said a coalition of tech organizations and trade groups, including the Computer and Communications Industry Association, which counts Alphabet, Apple, Amazon and Meta among its members, in a joint statement dated Nov. 27.

In the statement, the tech representatives said they were concerned “about the direction of the current proposals to regulate” generative AI systems and said the EU’s proposals “do not take into account the complexity of the AI value chain.”   

European lawmakers are also in disagreement over how to govern the nascent technology with EU member states France, Germany and Italy understood to be in favor of light touch regulation for developers of generative AI, according to sources close to the negotiations. 

In response, music executives are making a final pitch to legislators to ensure that AI companies respect copyright laws and strengthen existing protections against the unlawful use of music in training AI systems.  

Helen Smith, executive chair of European independent labels group IMPALA, tells Billboard that the inclusion of “meaningful transparency and record keeping obligations” in the final legislation is a “must for creators and rightsholders” if they are to be able to effectively engage in licensing negotiations.

In a letter sent to EU ambassadors last week, Björn Ulvaeus, founder member of ABBA and president of CISAC, the international trade organization for copyright collecting societies, warned policymakers that “without the right provisions requiring transparency, the rights of the creator to authorise and get paid for use of their works will be undermined and impossible to implement.”

The European Composer and Songwriter Alliance (ECSA), International Federation of Musicians (FIM) and International Artist Organisation (IAO) are also calling for guarantees that the rights of their members are respected.

If legislators fail to reach a compromise agreement at Wednesday’s fifth and planned-to-be-final negotiating session on the AI Act, there are a number of possible outcomes, including further ‘trologue’ talks the following week. If a deal doesn’t happen this month, however, there is the very real risk that the AI Act won’t be passed before the European parliamentary elections take place in June.

If that happens, a new parliament could theoretically scrap the bill altogether, although executives closely monitoring events in Brussels, the de facto capital of the European Union, say that is unlikely to happen and that there is strong political will from all sides to find a resolution before the end of the year when the current Spain-led presidency of the EU Council ends.

Because the AI Act is a regulation and not a directive — such as the equally divisive and just-as-fiercely-lobbied 2019 EU Copyright Directive — it would pass directly into law in all 27 EU member states, although only once it has been fully approved by the different branches of the European government via a final vote and officially entered into force (the exact timeframe of which could be determined in negotiations, but could take up to three years). 

In that instance, the act’s regulations will apply to any company that operates in the European Union, regardless of where they are based. Just as significant, if passed, the act will provide a world-first legislative model to other governments and international jurisdictions looking to draft their own laws on the use of artificial intelligence.

“It is important to get this right,” says IMPALA’s Smith, “and seize the opportunity to set a proper framework around these [generative AI] models.”