Hey, want to know whether your music is being illegally used to train AI? Well, now you can. Kind of. Thanks to a new tool that has been launched by the Atlantic magazine.
You may recall that a week or two back the US-based mag revealed that there are four giant datasets that are being shared within the tech community to train their AI models. Now with a simple search on a tool the Atlantic are calling AI Watchdog, you can see whether the creative piece that you slaved hours and hours over is included in these massive tranches of content.
Of course, the presence of your album, song or book in there isn’t proof in itself it has been illegally fed to AI. “AI companies may omit certain works when training, so the presence of a work in a dataset is not definitive proof that it was used,” said The Atlantic.
“Companies often use multiple datasets in training, so the absence of a given work is also not proof that it hasn't been used. Note that some datasets contain multiple copies of certain works.”
As revealed by The Atlantic two weeks ago, one of those datasets contains 12 million tracks, another has 9 million. There are two smaller ones, including one that originated as the Free Music Archive, which was started by New Jersey radio station WFMU as a service for fans back in 2009, during the increasingly distant era of tech idealism.
The question is, of course, once you’ve found your work on one of these datasets, what can you do about it? Sadly, the answer - unless you have the resources to take on the tech giants in court – is likely to be ‘not very much at the moment’.
When the head of the Free Music Archive Hessel Van Oorschot sent a letter alleging that Google had been training AI on the contents of its archive, their response was in his words “a big middle finger”. Their letter argued that “we believe everyone benefits from a vibrant content ecosystem.”
Change is only likely to come through effective legislation in both Europe and the States. Don’t hold your breath on that. In the meantime click here to see if your tracks can be found on The Atlantic’s AI Watchdog.
Search the Atlantic's AI Watchdog here>>
