Google and other Web 2.0 companies have a problem: like it or not, profit and liability go together when it comes to delivering content. Maybe opening up the algorithms will provide a solution.
Neutrality and the Common Carriage Bargain
At the heart of the net neutrality debate is an old bargain called common carriage [Christian Sandvig link to a small PDF]:
Common carriage is a common law legal concept that may date to the Roman empire… In brief, a common carrier is a private party offering transport or communication services who is subject to special public duties in return for legal benefits. The chief obligation of the common carrier is nondiscrimination—it must undertake to carry all people indiscriminately. (This is of course the center of the network neutrality debate.) Common carriers include railroads, taxis, airplanes, and telephones.
In exchange for this burden of nondiscrimination, common carriers have received a number of benefits: chiefly, liability protection. As common carriers can have no interest in the content that they carry, they are not liable for transporting stolen property—you can’t sue the phone company for copyright infringement if a telephone is used to read aloud a copyrighted work. Carriers may also not be liable for any other illegal content: offensive messages, indecent messages, or death threats.
The bargain is a problem for Internet Service Providers who want to carry out content-based traffic shaping to make the most of their investment in networks but who don't want to be liable for illegal or offensive content passing through their systems. Profit and liability go together: ISPs who argue that "we can't be responsible to the law for the content we deliver, but we have to be responsible to our shareholders for the content we deliver" have an obvious consistency problem.
Common Carriage 2.0
As the Web has morphed into Web 2.0, openness is less about filtering at the level of TCP/IP packets and more about what the major platform owners do. Peer to peer traffic has fallen off dramatically in the last couple of years, replaced by video streaming from YouTube and file sharing sites such as Rapidshare [link]. What responsibilities does Google have for the content of YouTube videos? Is Rapidshare a common carrier? And is there a role for search neutrality as well as net neutrality?
Here is where the common carriage bargain comes in. Google's advertising-driven business model is based on matching advertisements to content. As soon as Google started down this path it could no longer shrug its shoulders at regulators and say "who me? I don't know what's on YouTube. I'm just the carrier." To make money it built algorithms that know about the nature of the content, and the more it wants to make money the more those algorithms will know, and the copyright owners will have Google by the short-and-curlies.
Google's response has been to accept liability to preserve its advertising revenue. And where does that revenue come from? Yes, it's Ye Olde Media Companies [September 2008 link]:
McKinsey Quarterly: Will the Internet bring down barriers, making markets more democratic?
Eric Schmidt: I would like to tell you that the Internet has created such a level playing field that the long tail is absolutely the place to be—that there's so much differentiation, there’s so much diversity, so many new voices. Unfortunately, that’s not the case. What really happens is something called a power law, with the property that a small number of things are very highly concentrated and most other things have relatively little volume. Virtually all of the new network markets follow this law.
So, while the tail is very interesting, the vast majority of revenue remains in the head. And this is a lesson that businesses have to learn. While you can have a long tail strategy, you better have a head, because that's where all the revenue is.
And, in fact, it's probable that the Internet will lead to larger blockbusters and more concentration of brands. Which, again, doesn’t make sense to most people, because it’s a larger distribution medium. But when you get everybody together they still like to have one superstar. It's no longer a US superstar, it's a global superstar. So that means global brands, global businesses, global sports figures, global celebrities, global scandals, global politicians.
So, we love the long tail, but we make most of our revenue in the head, because of the math of the power law.
Last month Google followed through by partnering with Sony and Universal Music Group to launch Vevo, the YouTube-based site for music videos (currently available in the US, Canada, and Japan only). The videos on the site are encoded to prevent downloading. They are surrounded by ads. They are removed from the YouTube APIs so that third-party applications can't access them [link]. It's a closed site based on closed technology – so much for Google's commitment to openness as announced a week or two later [link]. But then, Google had little choice, given where its advertising money comes from.
The Vevo venture is extreme, but the same principle applies to other material on YouTube: the more Google extends its advertising throughout the site, the more it accepts the role of content cop, tracking down and removing material that is accused of breaking, the terms of copyright.
The issue goes beyond YouTube: here is Google's claim that it can't be held to blame if its image search algorithms give top ranking to an offensive picture of Michelle Obama:
Sometimes Google search results from the Internet can include disturbing content, even from innocuous queries. We assure you that the views expressed by such sites are not in any way endorsed by Google.
Search engines are a reflection of the content and information that is available on the Internet. A site's ranking in Google's search results relies heavily on computer algorithms using thousands of factors to calculate a page's relevance to a given query.
Or, "what me guv? I'm just the postman." But obviously Google can and does filter its image results by other criteria – it has a "safe search" option that excludes pornography but does not exclude racist images – and those criteria themselves are not chosen algorithmically.
Is Openness the Solution?
The advertising model is currently driving all before it when it comes to the Web, but while Google, Facebook and others will try to play both sides of the game, in the end you can't be both a common carrier and a successful ad-driven company.
Google's defense of "it's not me, its my algorithms" will be challenged at some point, and Google will either have to take responsibility for the content it delivers, or forego revenue. But maybe there is a third option, which is to open up the algorithms themselves. Right now, Google expects us to take on trust what it can and cannot do, but that defense won't last for long, particularly because of the barriers to entry for large-scale search and content hosting. One bargain it may be able to strike would be to keep both its
common carrier role and its ability to make money off advertising, but to lower the barriers to entry for competitors by opening up its code and its data centre architecture. Despite its vaunted commitment to openness, Google won't do this without a struggle, but maybe it will if it provides a way out of the common carrier dilemma.
Your covering a lot of ground here, Tom. I’m not convinced that Common Carrier is a fundamental principle that is required to grease the wheels of commerce. Common Carrier sounds like rhetorical lipstick for what is really government rent seeking.
In a system with economic freedom and rule of law that allows individuals to seek compensation for negligence, is Common Carrier really required? Was liability a real problem for carriage operators in ancient rome? Were Roman regulators actually trying to stomp the scourge of discrimination?
When it comes to net neutrality I do have some sympathy for the ISPs. I think part of the problem ISPs face is accurately planning for cost and capacity with the flat rate pricing packages that consumers prefer. Individuals that use their peak capacity 24hrs per day negatively impact the capacity of others. It is like the episode of the Simpson’s where Homer is thrown out of the All-You-Can-Eat restaurant. Homer is bad news for that business model.
There are certainly some cases where ISPs try to protect their own revenue models by preventing or limiting VoIP traffic or video downloads. One would hope that these are temporary blips that will be removed as customers voice their displeasure.
I also think the Copyright issue will iron itself out. We are in a transitional stage right now. Historically I’m sure copyright holders of music and movie rights struggled with the prospect of allowing music to be played on the radio and movies to be rented from Blockbuster. I’m positive that Google is determined to strike a balance that allows copyright holders to benefit from their content being available on YouTube.
If not, we will always have crazed caterpillars to entertain us 🙂
I think it would be great if we (society) reached some sort of “common carrier” accommodation with ISPs, search engines, and content hosts (eg: YouTube) because it would be of benefit to everyone.
But I disagree with you that responsibility (and liability) DOES vary inversely with the degree of control that the provider exerts. I think perhaps it SHOULD, but that’s not the same thing.
I fear that one likely resolution will be for governments to pass laws declaring that ISPs, search engines and content hosts are NOT liable for what they host, without including a concomitant responsibility for evenhandedness as with common carriage. That would have the effect of cementing in place large monopolies with massive powers to pick and choose winners and losers in fields dependent on these services, but I fear we (society) might make just such an error unless are are vigilant about avoiding it.
I have some sympathy for ISPs also, in that they are sometimes expected to be net-neutral and yet to be responsible for the content they provide, which is the worst of both worlds from their point of view. It would be interesting if competitive forces were enough to make common carrier irrelevant – FedEx managed to challenge the US Post Office after all – and that’s partly why I thought opening code up may help. If there is enough competition, then perhaps we don’t need as a society to make the common carrier bargain.
I have hopes that there are strong enough interests to push for this kind of accommodation. But it’s just a guess of course and you may well be right; if ISPs have a loud enough voice we may end up with exactly what you fear.
Tom, many of these issues were actually addressed in one form or another in the past decade, but it’s necessary to have followed the law/policy debate in detail to know it. One of the very weird experiences to me in the last few years is the effects of the recent anti-ISP politics – it’s the most literally Orwellian thing I’ve personally ever seen, in the way the politics simply was rewritten due to all the big-money interest involved (now I’m not saying it’s the most Orwellian ever globally, but rather, that I’ve experienced directly).
Anyway, the two big concepts you need to know about are “CDA section 230” (no liability for user-generated content) and “DMCA section 512 take-down” (no copyright liability if ISP’s take-down on notice). Plus the various legal cases involved. The site http://www.cybertelecom.org is a good place for an overview of this material, see especially http://www.cybertelecom.org/ip/dmca.htm
So – “ISPs who argue that “we can’t be responsible to the law for the content we deliver, but we have to be responsible to our shareholders for the content we deliver” have an obvious consistency problem.” – not really. The ISP’s got the law to exempt them from liability here many years ago. And the sort of network management at issue for ISPs are very different from the copyright issues. The hot issue right now is not ISPs, but platform-owners, who are fighting the idea that they have any sort of pro-active copyright obligation. And so far, they’ve been winning.
“As soon as Google started down this path it could no longer shrug its shoulders at regulators and say “who me? I don’t know what’s on YouTube. I’m just the carrier.” To make money it built algorithms that know about the nature of the content, and the more it wants to make money the more those algorithms will know, and the copyright owners will have Google by the short-and-curlies.”
An appealing chain of logic, but the flaw is in overestimating what the algorithms “know”. Google can say “The algorithm knows what this is about in a vague overall sense, but that’s nothing like knowing if it’s copyright infringement.”.
“Google’s response has been to accept liability to preserve its advertising revenue.”
Huh? I think you’re talking about “DMCA section 512 take-down” here. That pre-dated Google.
“the more Google extends its advertising throughout the site, the more it accepts the role of content cop, tracking down and removing material that is accused of breaking, the terms of copyright.”
You’re reasoning from an abstraction here – the legal details don’t follow.
“The advertising model is currently driving all before it when it comes to the Web, but while Google, Facebook and others will try to play both sides of the game, in the end you can’t be both a common carrier and a successful ad-driven company.”
Why not? It’s worked fine so far. Don’t argue “It’s logically inconsistent!!!”. That’s only even true in some sort of ultimate limit of perfect AI, which isn’t the case in the real world.
“Google’s defense of “it’s not me, its my algorithms” will be challenged at some point, and Google will either have to take responsibility for the content it delivers, or forego revenue.”
Not at all. I’d say they have a great defense in “My algorithms don’t do what you want them to do”.
“Right now, Google expects us to take on trust what it can and cannot do, but that defense won’t last for long, particularly because of the barriers to entry for large-scale search and content hosting.”
I don’t think this follows. And anything that’s examined as part of a legal case will be under very strong protective orders (Google is quite tough on this general aspect of legal cases).
I sympathize with what you’re trying to examine, but I fear you’ve been somewhat led astray by all the misinformation floating around on these topics. Anything you read about it nowadays is likely 99% nonsense or more.
Thanks for these valuable comment Seth. I am certainly not as up on the law as you are, and would not pretend to be, and your links and observations give me a lot to think about.
The recent developments that stand out are commercial practices rather than legal cases, although the two are obviously linked. When it comes to music videos and other copyrighted material on YouTube, for example, Google has had a lot of pressure to go beyond DMCA take-down to policing up front and identifying copyrighted material more pro-actively (but I don’t know what CDA says about this).
I may be overestimating what Google’s algorithms know, but I’d argue that (a) Google’s own words about what the algorithms can and cannot do are hardly disinterested and Sony is not likely to take Google’s own words unchallenged, and (b) to some extent it’s a matter of Google’s priorities rather than their limitations – face recognition in videos, for example, I’m sure Google could make a stab at if it’s worth their while. So Google is getting threatened by the major media outlets even without it going to court, and I think Vevo is Google’s attempt to accommodate those interests, and signals that the two groups will find ways to co-operate nicely in the long term.
You share Michael Chermside’s scepticism about whether advertising and common carriage can coexist. I’m usually a cynic myself, but I think there will be many special cases where these two get pitched against each other, and where Google may have to open up in order to answer the questions about how they do and can track and identify content. If advertising and common carriage can coexist, I just think that Google may have to pay a price and that price will be transparency (although what you say about protective orders may force me to rethink that). Not sure who will go after them hard enough – the EC perhaps? – but I think someone will sooner or later, and that when push comes to shove in a couple of years “My algorithms don’t do what you want them to do” will be challenged.