What Music And Journalism Tells Us About AI And Ourselves
AI is immensely powerful. Particularly generative AI. Sure, it has limitations, but it can do things our culture, law, and society are unprepared for. This isn’t just a case of ethics, either, as AI has the potential to ruin entire vital industries. As such, how industries react to AI today is crucial to securing their future. But it also tells us how these industries view us as consumers. There is no better example of this than the polarising actions of the music industry and journalism.
The music industry is regarded as quite cut-throat. Record labels and producers are renowned for pushing even the biggest artists into seriously horrific contracts. If you don’t believe me, look at Taylor Swift or Prince. This industry can be the personification of corporate greed, bleeding every asset dry.
You’d think this would make them adopt AI technology incredibly rapidly. After all, thanks to algorithm-based streaming services, music is getting more generic and formulaic. In theory, it would be effortless to get an AI to pump out thousands of songs with the statistical hallmarks of a №1. Moreover, these AIs wouldn’t need to be paid, wouldn’t dispute contracts, and could even play multiple “live” shows a night.
But, despite these advantages, the music industry isn’t adopting AI at all. In fact, it is actively protecting its artists from AI. A few days ago, Sony Music, Universal Music Group and Warner Records all sued AI music generators Suno and Udio for copyright infringement on an “almost unimaginable scale”.
Why? Well, Suno and Udio extensively used these record labels’ music to train their AI. Like all AI firms, they claim this use is legitimate under “fair use”, which allows copyrighted works to be used for commentary, satire, news or if the use is transformative. However, the record labels have argued that “The use here is far from transformative, as there is no functional purpose for… [the] AI model to ingest the Copyrighted Recordings other than to spit out new, competing music files” and that the “motive is brazenly commercial and threatens to displace the genuine human artistry that is at the heart of copyright protection.” Indeed, it is reportedly very easy to get both Suno and Udio to create works that sound almost identical to artists like ABBA.
Being titans in their industry, this lawsuit is going for the jugular and is asking for $150,000 per work, which resembles their copyrighted material. As Suno alone has raised $125 million and has already had over 10 million customers, if the court agrees with the record labels, they could have to pay billions of dollars in compensation.
These record labels aren’t messing around. They understand that their industry operates on human-human connections and trust. Flooding this industry with AI knock-offs would drown these valuable connections out, sniffle innovation, and grind the entire industry to a halt. So, they are viciously protecting their artists and the industry at large.
Compare that to journalism. An industry that is, arguably, entirely based on human-to-human connections and trust, and is far less formulaic and detailed-focused, making it harder for AI to replicate. There is zero incentive for big journalistic publications to extensively use generative AI over its highly trained workforce.
Yet, they all seem to be queuing up to sign deals with OpenAI to train ChatGPT on their past and present content. Vox, Politico, Business Insider, the Financial Times, Investopedia, the Wall Street Journal, New York Times, The Times, The Sunday Times, The Sun, and the Atlantic and many more have all signed such deals. This will enable ChatGPT to perfectly emulate their editorial style and subject matter. What’s more, none of these publications consulted their journalists or journalist unions before signing away their work like this. In fact, most found out about these days from the news rather than from within the company.
So, not only are these publications refusing to safeguard their human talent from AI, but they are actually actively feeding their work into the AI machine. I’m sure these deals are relatively lucrative, but I can also guarantee these publications are busily engaging with OpenAI to replace their journalists with ChatGPT. Why else would they be okay giving the AI the data it needs to replicate their unique editorial style?
This demonstrates how restrictive and motivated journalism has become in recent years. These companies don’t respect their journalists’ beliefs, ability to investigate, creativity, ability to appeal to others, human voice, or workers’ rights.
But imagine what will happen to journalism once these publications adopt ChatGPT? For one, their articles will get less accurate, as ChatGPT still suffers from hallucinating fake facts, and plenty of evidence suggests this can’t be fixed by training it on more data. But it also means they can pump out thousands of these potentially not-so-accurate articles daily on any topic, drowning out real human voices, human stories, and genuine connections.
These attitudes of journalistic publications and record labels are not new. AI is just highlighting how these industries treat their talent. Publications have become less profitable over the years, and they are shutting down left, right and centre. Meanwhile, the music industry is growing by around 10% annually, and the number of new record labels is booming.
So, what does all of this tell us about AI and ourselves? In short, it shows that creative industries can go from strength to strength when they respect human workers and human connection. But, when they value cheap, vast, uncreative, low-quality work that doesn’t talk to our inner being, the industry shrinks. We are human; we are social. AI can be used to augment and enhance these connections, but it cannot replace them.
Thanks for reading! Content like this doesn’t happen without your support. So, if you want to see more like this, don’t forget to Subscribe and help get the word out by hitting the share button below.
Sources: The Guardian, BBC, Tech Crunch, Mashable, Reuter Institute, Statista, The Drum