Music to a Lawyer’s Ears
By Gordon Hopkins
Like radio, television and the internet before it, AI (Artificial Intelligence) will change our world in countless ways, many of which we will be unable to predict or even truly comprehend.
One thing we can predict, however, is that AI will bring many, many, many lawsuits.
Generative AI “learns” by taking in content, be it visual, written or music, created by living flesh-and-blood humans. It then creates it’s own content, using what it has “learned.” I’ve already explained in previous columns why what AI does isn’t actually “learning,” which is why “learn” is in quotation marks. The larger the pool of data, of content, it has to draw from, the more convincing the results. That means AI has an insatiable appetite for content. An AI can’t create a song if it doesn’t already have access to hundreds of other songs.
And there lies the legal rub, because most of that original content the AI is using was created by artists and is copyrighted.
Some industries are quicker than others to take on the incursion of new technologies. The music industry, for example, has a long history of fighting for it’s intellectual property rights tooth and nail. Such as the legal fights that took down Napster and other file-sharing services that were used to skirt copyright.
Sony Music has fired a warning shot over the bow of AI. A letter sent by Sony to more than 700 AI companies warned they better not be using their music without permission, “We will take any necessary steps to prevent the infringement or other violations of our artists’ and songwriters’ creative works and rights.”
While some AI companies are trying to work with content creators to avoid the tsunami of legal actions, but others are drawing an electronic line in the digital sand, insisting it falls under “fair use,” which is the concept that copyrighted material can be used without permission under certain circumstances. Some countries, although not the U.S., have data mining copyright exceptions. On the other hand, some countries, also not the U.S., allow copyright holders to opt out of the data mining exception.
The bottom line is, current U.S. law probably isn’t sufficient to address AI in an meaningful way. At least, as a non-lawyer, that’s how I see it. This is legally untrodden ground. So, the lawmakers will have to make some new laws.
So, the fight begins. It will be long. It will be ugly. Above all, it will be expensive. My suspicion is, at the end of the day, the only real winners will be the lawyers.
