Who is copyright protecting?
Readers of this blog who are under 30 may not even know what a cassette tape is! But when they appeared in the 1970s, they were a revolution allowing music to be easily copied from legal media such as vinyl records or other cassettes. Arguably, the main purpose of buying a blank cassette and having a cassette recorder was to copy music, but, of course, there were also other purposes such as recording your own music, lectures, or letters. Even the nascent home computer industry was starting to use cassette tapes as a cheap alternative to computer disks.
The New York Times wrote in 1985 about the challenges of copyright in the electronic age. Little could they have imagined the challenges of the digital age some 20 years later! The trigger for the article was a bill before Congress at the time to charge a levy on blank cassette tapes. Music executives worried about their rights, and revenue, being lost.
Like the US, such a solution was suggested in many jurisdictions. Despite the protests of legal users, many countries still have a small levy on blank media, although it is more likely a CD these days, and distribute it imperfectly to artists.
More importantly, the battle over cassettes further matured a decades-long licensing approach that is now relatively universal. Regardless of how you source your music, if you are sharing it you are required to pay a license fee. Depending on where you live, retail shops, barbers and offices all have a responsibility to pay a fee even if they are playing media that they own outright, streaming or just have the volume turned up on their radio.
Music executives tried to extend their license protection efforts against Napster in the early noughties. Most readers will remember Napster as freeing up music copying on a massive scale through illegal peer to peer file sharing. It was shut down after just two years as the music industry fought back in the courts. But the disaggregation of music couldn’t be tamed, first iTunes and then ultimately Spotify revolutionised the music industry and it still has not found its financial footing.
What Napster did to music, search engines and social media repeated with newspapers while mostly staying on the right side of the law. Even films and books haven’t gone unscathed with a combination of streaming, digital reading and online retailing transforming the marketplace.
Now the ecosystem around just about all digitised material is again being reimagined with Generative AI. As with the cassettes of the 1970s, the technology is moving faster than the decisions we can make as societies seeking to balance the needs of producers and consumers of content.
While most people accept the importance of defending IP owners, it is far from clear who or what is being protected. When record labels asserted their rights, they weren’t protecting revenue on behalf of artists, the majority of whom have always received a pittance for their albums. Executives were protecting their own lucrative contracts and left most artists to rely instead, as they still do today, on live performances. Nor were they protecting the identity and provenance of the music given audiences knew exactly who they were listening to.
For newspapers today negotiating with search engines and social media, it isn’t about protecting journalists who are already a threatened species, it is about advertising revenue. Newspaper proprietors had already demonstrated a willingness to exchange jobs for productivity and revenue through the industrial battles waged across many mastheads globally in the 1980s and 1990s.
All too often, discussions on protecting IP have been about protecting lazy legacy businesses against disruption rather than ensuring that content creators are both recognised and rewarded for their labour even as business models transform.
Understanding this is important as Generative AI takes past electronic and digital disruption to a new level of disaggregation by converting virtually the entire body of the Internet and more into weighted tokens inside Large Language Models (LLMs). In doing so, the provenance and quality of the content itself can easily be lost.
Each person involved in creating content deserves to be recognised for their contribution, have some level of control of its application and receive a fair share of any profits earned by its use. It should not be a given that any existing or legacy business model should be protected as a channel to market, but the role of each intermediary should be recognised. Journalists and writers benefit from editors; publishers commission content that might not otherwise be created; and distributers get content to audiences who don’t otherwise know of its existence.
As is often the case, new technology could be part of the answer with emerging work on digital watermarking demonstrating provenance even after disaggregation into LLMs. Given that most content is sourced from the Internet, we may even have finally have a compelling use for the Semantic Web which associates machine readable metadata with the web page content.
None of this, though, will be useful unless regulators can find a licensing regime where end-users are prepared to pay a fee to use Generative AI which is distributed back to content producers. But given the precedent of music in the electronic age, it is likely that the solution is an evolution of past levies on blank media rather than something more revolutionary.