Back, Bartok, Schubert, Puccini and Wagner wrote their compositions hundreds of years ago before the idea of copyright was even a glint in an intellectual property attorney’s eye, so they’re in the public domain. But are they? It turns out it’s getting tougher and tougher to post the music of the classical greats by a performer today without receiving a notice of a copyright violation from some company somewhere. The problem is that most of them are fraudulent claims, but that doesn’t keep them from happening and making a classical performer’s life a little more miserable.
One case is pianist James Rhodes who put an innocent enough video of one of his Bach performances on his Facebook page and received a copyright violation notice from Sony Music Global claiming that they owned 47 seconds of his performance. The composer has been dead for 300 years, but that didn’t stop Sony from claiming ownership.
The German music professor Ulrich Kaiser found that it was nearly impossible to post anything by any of the major composers because a number of companies claimed they owned the catalogs of these classical music giants.
So here’s what probably happened. Sony Music et al do control the recordings of many performers that have recorded compositions by the great classical composers. YouTube’s ContentID and Facebook’s similar content filtering system spotted the similarities between the recorded versions owned by the record labels and the new indie content and flagged it, resulting in the copyright violation notice.
Essentially, these content systems are smart but not smart enough to know when the composition is out of copyright and in the public domain. As far as these systems are concerned, it’s as if the label’s versions were being used as source music behind some video. In other words, they can’t tell the difference between versions since they’re played so precisely. Bad AI!
The scary thing is that this can happen to text, photos or any other online content. I’ve been flagged for a copyright violation for writing on the same news item as someone else, and I’ve been flagged for quoting my own material as an excerpt. Or you could take a picture of a building but get flagged because someone else took the same picture from the same angle and posted it before you.
Bottom line, the algorithms used in these content identification systems must get smarter fast, or content creators are in for giant hassles in the future.