The Lawsuit That Could End The AI Industry
The Lawsuit That Could End The AI Industry
Can we finally hold OpenAI and alike to account?
Generative AI has a bit of a copyright issue. It takes an utterly vast amount of data to train these remarkable programs, too much for an AI company to create in-house. As such, they pirate copyright data from all over the internet, from social media posts, newspapers, books and even YouTube videos. Understandably, many artists and creators are furious at this, as they aren’t being compensated for their work being used in this way, and once these AI models are trained on their work, they can produce crap copies of it on a vast scale, potentially drowning the original artist out. So, how can we stop this deleterious widespread pillaging? Well, a YouTuber may have just found a way, and his lawsuit against these AI companies could actually end the AI industry as we know it.
You may ask how these AI companies can use copyrighted material in the first place. The answer is “fair use.” Fair use is a part of copyright law meant to enable third parties to comment on or analyse copyrighted material. As such, it permits a party to use a copyrighted work without the copyright owner’s permission for purposes such as criticism, comment, news reporting, teaching, scholarship, research or transformation into something new. This is how YouTube commentary channels can use copyrighted videos from films, TV shows, and video games.
AI companies like Nvidia and OpenAI have so far successfully claimed that their use of copyrighted material falls under fair use as they transform the material. As such, all copyright cases filed against AI companies have failed to bring charges. But, while this use technically falls under fair use, it doesn’t follow the ethics of fair use, as these AI programs enable others to easily replicate the copyrighted material they have used rather than commenting on or reporting on said material. Moreover, fair use was only meant for small players and not this scaled-up industrial use. As such, many artists and creators have suggested copyright law needs to be altered to better suit modern needs.
But this is where David Millette, the owner of the YouTube Channel Legal Drive, comes in. You see, he is bringing two separate lawsuits against OpenAI and Nvidia for using his copyrighted material to train their models. However, he isn’t accusing them of copyright infringement. Instead, he accuses them of “unjust enrichment and competition,” an angle no one has taken before.
Unjust enrichment is defined as “benefit by chance, mistake or another’s misfortune for which the one enriched has not paid or worked and morally and ethically should not keep. If the money or property received rightly should have been delivered or belonged to another, then the party enriched must make restitution to the rightful owner.” For a court to find there has been unjust enrichment, there must be three things. Firstly, the other party was enriched. Secondly, this happened at that party’s expense. Finally, that was against equity and good conscience to permit the other party to retain what is sought to be recovered. If all three of these are found to be true, then the court can order the other party to compensate the party.
Well, it’s difficult to see how OpenAI and Nvidia, who are now racking in billions of dollars from their AI models, don’t tick every single one of these points. As such, many feel these lawsuits will be slam dunks for Millette. If this happens, it will set a precedent, and any generated AI company that has scrapped copyrighted material will have to dish out a truly huge amount of compensation, potentially to hundreds of thousands, maybe even millions of claimants.
But this could kill the AI industry.
OpenAI has been very transparent that AI models like ChatGPT would be impossible without access to copyrighted material. If Millette’s cases fall in his favour and set a precedent of proper compensation for copyright holders, no matter how small, it could end the AI industry as we know it.
You see, generative AIs are not profitable at all.
Firstly, by the end of this year, the entire AI industry is set to have spent $150 billion on AI GPUs. Analysts have found that the costs of installing and running these GPUs and generating enough revenue to maintain this new infrastructure would mean the AI industry needs to generate an annual revenue of $600 billion. This simply isn’t possible. MIT’s Daron Acemoglu found AI will only improve US productivity by 0.5% over the next decade and will only increase US GDP by 0.9%. This predicted GDP increase amounts to only $2 billion, making that $600 billion revenue increase utterly impossible. To top that off, OpenAI is on course to file a net loss of $5 billion by the end of the year, with many predicting it will go bankrupt.
In other words, the AI industry has already made its own coffin. It doesn’t have the revenue to keep itself afloat as it is. If a precedent is set to adequately compensate these copyright holders, no matter how slight, it would dramatically amplify these revenue issues and be the final and definite nail in this coffin.
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: Toms hardware, Dictionary.law, The Guardian, Will Lockett, Copyright Alliance