Alternative Compensation Systems

With every passing day, online music downloading becomes more prevalent and industry countermeasures become more odious. What if there were a compromise that paid artists while letting you get music however you wanted? This is the idea behind "alternative compensation systems" and "compulsory licensing".

Here's the proposal in a nutshell: Some group of people pay a small fee (like a couple dollars a month). In return, they can download whatever they want, however they want. We track what is downloaded and then distribute the money received, in proportion, to the people responsible for the songs. Everybody wins: users get all the music they want, software developers can continue innovating, and the industry gets paid.

Several people have prepared more detailed proposals. Terry Fisher of the Berkman for Center of Internet and Society at Harvard Law School brought together a group of the interested parties to hash out the details of such a system and think about ways to deal with the problems. We discuss these things on the moderated list acs@pholist.org (subscribe, archive). I will try to keep this page up-to-date with the latest proposals, issues, and solutions as discussed on that list.

Comments may go to acs@pholist.org (moderated list) or acs@aaronsw.com (my personal email).

Consensus Proposal

This is the proposal that most seem to agree on and that best solves all the known issues.

People with broadband Internet connections pay an extra $3-$5 on their bill (the amount is calculated by determining how much money is lost by Internet downloads and dividing it by the number of Internet connections), which is given to the operators of the system ("MiniPay"). (Optionally, a similar charge can be added to other Internet connections, blank media, burners, computers, etc.) In return, they receive a "blinded signature token" that allows them to tell MiniPay what they're listening to without cheating (the token can only be used once) and without knowing who submitted it (the ISP was "blinded" when it gave it out, so they don't which they gave to who). The token is sent to them automatically and electronically and is held onto by their MP3 player (e.g. iTunes).

Copyright holders, when creating new songs and movies, submit a copy to MiniPay. MiniPay uses an audio/video "fingerprinting" technology to calculate a short number which identifies the song. They store this with information about the copyright holder, so they know who to distribute payments to later.

Now people surf the Internet, downloading whatever songs they like, however they like, and listening to them. Their MP3 player keeps track of how many times they listen to each song (identified by using the same fingerprinting technology as MiniPay). If they use portable devices, the devices keep track and submit the information back to the MP3 player. If multiple computers share an Internet connection, the computers also share playcount data. At the end of the month, the computer collects all the data and anonymously submits it to MiniPay with the token.

MiniPay divides up however much the token is worth (the $3-$5 charge less operating expenses) in proportion to how much each song is played. Then it looks at who holds the copyright to each song and deposits the money in their account.

Issues

Artist Abuse

The copyright holder? Won't that be the record labels? And don't those guys always cheat the artists?

The artist, at least initially, holds the copyright to their work. In the past they were forced to give it up to record companies, because those companies effectively controlled the distribution channels. But anyone can distribute songs in this manner over the Internet, so artists will be able to choose to market songs themselves in this system. (If they think the record companies provide a useful service, then they are of course welcome to give up their copyrights if they choose to.)

Eugene's Problem

People could send inaccurate playcount data to MiniPay. For example, if they like the NRA, they could claim they only listened to the NRA and thus easily make a donation to them with (seemingly) no cost on their part. (Eugene Volokh)

Yes, this is one difficult problem, but there are some ways to mitigate it. First, ads and publicity can be used to shame the NRA and discourage people from doing this. Second, MiniPay could cap the percentage of any single token that goes to a single copyright holder (proposed by Terry Fisher). Third, a variety of technical means could possibly be used to make faking playcount data difficult for the average user.

The hope is that such gaming would be below a certain acceptable level. It is important to keep in mind that similar current compensation systems (e.g. the Nielsen ratings) are themselves highly inaccurate and face similar problems. It seems likely that even with this kind of gaming, the system would be more accurate than those.

Couldn't I start a service that offered to buy tokens for a little under their actual value? (Seth Schoen) I could use this to launder money. (Ernest Miller)

Essentially, this is a special case of Eugene's problem. But how would these organizations be able to tell real tokens from fake tokens? I (or MiniPay) could sell them dozesn of fake tokens (indistiguishable from real tokens to anyone but MiniPay) and make a lot of money, while bankrupting then. MiniPay could also see which accounts these tokens were deposited into and then shut them down and perhaps even punish those who opened them.

Viruses

What stops me from writing a virus that submits falsified playcount data to MiniPay, claiming everyone has listened to my song a lot? Or, if you close my account if I do this, then I'll claim everyone has listened to my enemy's song a lot.

Yeah, this is a hard problem. We could set up honeypots to catch samples of the virus and see what they do and try to negate it. But this isn't possible in all instances because the virus could throw some data away. Perhaps the best thing is to include some quickly-updated anti-virus software with all playcount collecting software so that any such virus can be quickly contained.

Auditing

How do we know MiniPay won't lie about who gets what? (Mark Bernstein)

We could have several MiniPay-like organizations who are unlikely to collude. Users would send their information to each, and each would do the payout calculations. If there were any large discrepencies, we could send in people to physically investigate.

Open Issues

Calculating Quantities

How do we determine how much to pay out (and thus how much to charge)? We want to pay back whatever sales are lost due to the Internet, but that's very difficult to calculate. First, not all sales lost are the Internet's fault (sales could be dropping because nobody likes the music being put out, the price points, or some other reason). Second, not all Internet downloads are lost sales (many people download songs because it's easier/cheaper than buying). It gets worse when the Internet becomes the dominant medium for downloading songs, since we have nothing to compare against. This is still an open problem.

Derivative Works

People will inevitably create mashups and remixes, something which is illegal without permission under current copyright law. We could do several things:

  1. Keep it illegal.
  2. Allow it, but give the money to the original artists and not the remixer.
  3. Allow it, and give money to every artist involved (including the remixer).

1 doesn't seem practical (if we're not outlawing P2P, how will we stop it?). 2 seems unfair to remixers. 3 is unacceptable to many artists (they're getting money off of my work?!).

Even if we do decide that, then how do we decide what percentage of the work goes to each artist? We could simply measure quantities of the work ("40% Britney, 20% J.Lo") but what about more complicated remixes (e.g. a Britney-J.Lo mashup) or intangible contributions (how much does the remixer get? what about someone who adds animation? what if you set a poem to music?).

Some sort of arbitration body (like the Writers Guild of America has to decide who gets credit for collaboratively-written scripts) would make sense, but how would we pay for it?

Related links

My previous stories on compulsory licensing:

Peter Eckerlsey is working on a PhD thesis on this topic.