The copyright questions posed by generative AI technologies were in the spotlight in US Congress yesterday, with music-makers given the opportunity to set out their concerns and priorities before the House Judiciary Committee. Meanwhile, the former General Counsel of the US Copyright Office was asked to what extent copyright law can even address those concerns.

AI tools and technologies that can compose and produce original music have become big talking points of late within the music industry, of course, partly because said tools and technologies are becoming ever more sophisticated, and partly because of the general hype around generative AI caused in no small part by ChatGPT.

While neither the technologies nor the copyright concerns they pose are new, the need to address said concerns feels a lot more urgent. Hence the launch in March of the music industry-led Human Artistry Campaign and the eagerness of music-makers to speak at a session of the Judiciary Committee’s Subcommittee On Courts, Intellectual Property And The Internet with the title ‘Interoperability Of AI And Copyright Law’.

One key copyright question relates to what licences the makers and users of music-making AI tools need when they train said tools by crunching data connected to existing songs and recordings. Then there is the question of whether music created in that way should enjoy copyright protection.

Beyond straightforward copyright law queries, there has also been a demand for more transparency around what AI tools are being employed in the creation of content and what data sets are being used to train those tools.

And in more recent weeks – with the flurry of AI-created tracks that mimic the voices of pop stars – there has been increased debate over whether or not an artist can protect their voice, over and above any control they may or may not have via copyright when it comes to the training of AI technology with their past musical output.

Among the music-makers speaking in Congress yesterday was Dan Navarro, who is also involved in the aforementioned Human Artistry Campaign.

He ran the politicians through the key asks of that campaign, including that the makers and users of AI tools must get permission whenever they train those tools with existing music, and that lawmakers must not introduce any new copyright exceptions that remove that requirement.

He also stated that “copyright should only protect the unique value of human intellectual creativity. The copyright clause of the Constitution exists to incentivise humans to create – machines don’t need incentives”.

And, he added, “trustworthiness and transparency are essential to the success of AI and protection of creators. Without transparent AI, we will have no idea whether the inputs AI systems were trained on were licensed, leaving us no way to enforce our rights”.

Similar concerns and demands were expressed by composer and producer Ashley Irwin, who is also President of the Society Of Composers And Lyricists.

“Generative AI has been equipped using copyright-protected human-authored works and programmed to mimic those works without consent, compensation or credit”, he noted, and not only that, but “copyright information – ie metadata – has been removed during the ingestion process of these models”.

To overcome these concerns, the makers and users of generative AI tools must adhere to three fundamental principles, he stated. Those being consent, credit and compensation.

“Consent by creators for the use of their works in generative AI media”, he explained, “credit, wherever creators’ works are used, and compensation, at fair market rates, for the ingestion of any portion of human creators’ copyrighted works by AI generative machines and the subsequent output of new derivative works”.

Although the corporate interests of the music industry were not asked to speak at yesterday’s session, the bosses of the Recording Industry Association Of America and the US National Music Publishers Association – Mitch Glazier and David Israelite – penned an op-ed in Billboard earlier this week that was referenced in Congress. And it confirmed that, on AI issues, at the moment music-makers and music companies are more or less in agreement.

“Use of copyrighted works to train or develop AI must be subject to free-market licensing and authorisation from all rightsholders”, they said. “Creators and copyright owners must retain exclusive control over the ways their work is used. The moral invasion of AI engines that steal the core of a professional performer’s identity – the product of a lifetime’s hard work and dedication – without permission or pay cannot be tolerated”.

“This will require AI developers to ensure copyrighted training inputs are approved and licensed, including those used by pre-trained AIs they employ”, they went on.

“It means they need to keep thorough and transparent records of the creative works and likenesses used to train AI systems and how they were exploited. These obligations are nothing new, though – anyone who uses another creator’s work or a professional’s voice, image or likeness must already ensure they have the necessary rights and maintain the records to prove it”.

Committee members also asked lawyer Sy Damle – formerly General Counsel for the US Copyright Office – whether copyright law as it currently stands could address the concerns raised by others speaking at the session. In the main he believes it can, even if that means relying on the always tricky concept of fair use to balance the interests of copyright owners and technology makers.

“I think existing law is well suited to deal with all the questions we’ve been talking about today,” Damle said, according to Law360. “Congress had the wisdom in the 1976 [Copyright] Act and various amendments since then to build a technology-neutral flexible copyright regime, and I think it’s proven time and time again that no matter the new technology that comes along, the laws are able to adapt to them”.

That said, when it comes to artists protecting their voices, Damle agreed that that goes beyond the remit of copyright. “I think copyright law is too blunt of an instrument to deal with that”, he told the committee, “and it may be worth looking at other areas of law outside of copyright”. It seems likely that things like trademark and publicity rights are more relevant in that domain.

Following yesterday’s hearing, organisers of the Human Artistry Campaign said in a statement: “The more Congress learns about AI, the more members appreciate the moral and legal imperative to respect creators’ right to control how their work is used and for strong public policy to ensure professional performers’ voices and likenesses aren’t cloned and impersonated for commercial gain. We are grateful to all the members of the House Judiciary Committee for digging so deeply and seriously into these vital questions”.

‍