May 24, 2005
Will Automated News Writers Violate Copyright Law?
Lawyer, novelist, and aspiring screen writer Julie Hilden writing in FindLaw analyses a hypothetical not very distant future created by Robin Sloan and Matt Thompson where a hypothetical merger of Amazon and Google produces sophisticated software which reads all the news stories on the web and writes personalized news articles that synthesize the knowledge from many news sources.
In 2010, Googlezon wins its fight with Newsbotser by inventing a clever new technique that further tailors content to the user: "Googlezon's computers construct news stories dynamically, stripping sentences and facts from all content sources and recombining them. The computer writes a [personalized] news story for every user."
For the copyright holders who funded the writing of the original stories the problem arises when increasing numbers of readers no longer read the original stories. The New York Times or Washington Post get fewer visitors to their site or fewer subscribers for home delivery and hence less revenue from ads and subscribers.
So to give an example - mine, not Sloan and Thompson's - suppose the knowledge that Googlezon uses to customize content indicates that a particular user is very interested in international news. In putting together a news story even on a domestic happening, Googlezon could emphasize the international aspects - stripping from other sites (including blogs), say, only five sentences on the domestic happening, and twenty sentences on its international implications, to make a story.
Returning to Sloan and Thompson's predictions, in 2011, the New York Times and other media whose content is not customized to the user go the Supreme Court, "claiming that [Googlezon's] fact-stripping robots are a violation of copyright law."
But - according to Sloan and Thompson - the old media lose. As a result, they disappear - relegated to the status of newsletters for the elite and the elderly!)
Read the full article for Julie's discussion of the legal angles and how the hypothetical future legal case might turn out.
Consider how media firms might respond if no legal remedies are available. One response would be to reduce what gets put on the web. But even if the New York Times reduced how many articles it put on its web site the mythical future Googlezon company could still buy hard copies of daily NY Times newspapers, scan them into their computers, and use optical character recognition technology to convert the articles to text. Then their software could parse that text just as search engines read text on web pages. The knowledge extracted by the sophisticated parsing along with the knowledge extracted from all the other newspapers in the world could still be used to write news stories.
In one sense the Googlezon automated news story writer is not qualitiatively different from what already goes on. Talk show hosts read newspapers and then tell their listeners about what is going on. Have you ever watched C-SPAN where they read local newspapers from around the country? Or look at web logs which excerpt from various newspapers and other news sources.
The problem with the Googlezon idea is that once it becomes possible (and I think it is a matter of when, not if) this future tool will very efficiently separate the original gathering of information from the writing about it so thoroughly that news organizations that pay news gatherers (a.k.a. reporters) to go into the field to collect information will have a hard time generating much revenue to pay for their information gathering efforts.
I see this as part of a wider problem. Movie makers, music makers, software writers, and other content providers suffer economic losses from bootleggers who make copies of work which is copyrighted or in some other way owned as intellectual property. New technological developments continue to make copying easier. Faster internet connections make downloads easier. Cheap internet hosting services and the spread of the internet to countries with less intellectual property protection provide easy ways for bootleggers to make copyrighted material accessible for copying in violation of laws in many jurisdictions. Mass storage device capacities keep going up while their sizes shrink.
As the Googlezon example demonstrates, not all undermining of the value of legally protected material takes the form of violation of copyright law or of violation of other existing intellectual property law. Here is the root problem: New methods of information processing and information distribution effectively undermine old models for funding knowledge gathering and content production. The new methods of information processing, transmission, and storage may not - at least initially - provide a different set of incentives and mechanisms for doing as much content generation as is done today.
This wider problem extends even to the point of reducing the economic incentives for governments to fund basic research. When information flowed much more slowly basic research funded in one country which yielded economically valuable discoveries was most likely to go through commercialization where the basic discoveries were made. The value of propinquity between researchers and start-up companies was so high that, say, a Stanford professor who made a commercially valuable discovery most likely caused the formation of a company near Stanford to commercialize the knowledge from the discovery. But in the future the odds of a discovery in America first being commercialized in China or other countries will become much higher as the details about scientific discoveries propagate around the world more quickly via the web. When that happens many governments will see fewer incentives to pay for basic research. The resulting decrease in funding will be a loss for us all.
These problems are market failures. The concept of property has been extended from land and physical things into ideas, designs, and other intellectual creations that basically amount to patterns. This extension of property into the intellectual realm has caused a boom in efforts to generate useful knowledge and other products of the mind. But technological advances are increasing the cost of protecting existing intellectual property. The market for intellectual property increasingly fails due to the ease with which people can benefit from knowledge without financially contributing to its creation. Those technological trends in computing and communications show no sign of running out of steam. The solutions to this problem are not obvious to me.
Update: On the bright side advances in computing and communications technologies also lower the cost of knowledge creation and other forms of content creation. For each type of content we need to ask what is happening more quickly: Is "stealabilty" (for lack of a better term) going up faster or slower than costs of patterns production are dropping? Even if "stealability" is not advancing more rapidly than technological advances lower production costs we still suffer losses from a reduction in the amount of content generated as compared to an environment where more content use must be paid for.
Business models can be adjusted to respond to the new environment of rampant intellectual property theft. For example, rather than sell a shrinkwrap piece of software that does some function a business can sell the service of actually doing that function. Then the software can be loaded only on a server controlled by the business and customers can send in data to be transformed by the software or interact with the software over the web and pay per use. However, these adjustments will not restore all the incentives for knowledge production which are being lost due to easy copying and uncompensated reuse.
The solution to the Googlezon issue seems obvious:
Create a company that's devoted soley to creating sievable media. Make it convenient to do so, work with the precis-creators. And charge them for access to your corpus.
Can they get the data elsewhere? Maybe. But as traditional primary sources fade from view, your supply of raw data becomes ever more valuable. Your RELIABLE supply of raw data becomes moreso. The Old Media (MSM, whathaveyou) is rapidly cutting away its social credit, spending it unwisely. There's certainly a niche for what used to be considered "reporting." Gathering the facts, formatting them, putting them out for consumption, in exchange for pay.
That much is obvious.
For bootlegged movies? Try offering the audience something the bootlegger can't. (I'd suggest quality content, but that day is long past.) Theatres still have the lock on the communal audience experience that plays used to (and Lucas' whining about BitTorrent biting into his $150 mil weekend is pure mockery of reason). Offer other intangibles. Get creative.
The real problem is the ongoing belief that, somehow, the world owes markets continuing profits. Tell that to the buggy-whip manufacturers.
As soon as a raw data service sells the info to Googlezon and Googlezon puts it on the web then another competitor to Googlezon will read Googlezon's pages and then construct their own news stories to compete with Googlezon.
I do not see what moviemakers can offer than bootleggers can't steal.
Stealability sounds so immoral. How about ease of sampling? It also has one fewer syllable.
So what, you want to be able to rip off IP with a clear conscience?
Why should it be a matter of "conscience"?
Copyright law is fundamentally a compromise -- society (reluctantly) grants limited monopolies over certain ideas in order to provide incentives to create those ideas. And now, as you point out, technological change is astronomically increasing the costs of maintaining these sorts of monopolies.
Rather than tossing around loaded terms like "rip off" and "steal" and "theft," it might be more productive to start thinking harder about ways to compensate creators that don't involve copyright or grants of monopoly.
As I see it consciences are absolutely essential elements of a healthy free society.
So, what, is "conscience" hopelessly old fashioned? Should we perhaps genetically engineer our offspring to not have consciences? Or just engineer and teach them not to feel bad about taking the intellectual work of others without compensation?
I have been worrying about this problem for many years and have few suggestions to offer.
Greater government funding of science has to compete against demands by various interest groups for things that immediately benefit them. Plus, libertarians claim that government should have no part in funding science. So I do not expect government to do much to help with funding.
Plus, government is ill suited to choosing, say, artists or novelists or filmmakers or product developers to fund.
Generally I find that people who are happy to see the decay of protections of intellectual property have nothing to offer as a substitute except to tell other people to just work for free.
I agree with you on the role of conscience. But should I have a guilty conscience when I break the speed limit? When I fail to teach creationism in my Kansas biology class? Not all violations of the law should engender a guilty conscience.
But more to the point, not all people who agitate against the current state of copyright are motivated by the need to assuage a guilty conscience. Some of us look at the costs of enforcement and decide that they're not worth the corresponding benefits. Should we feel "guilty" about this? I don't see why.
I also agree with you that the government is a poor choice to fund art. And furthermore, I would never tell anyone to work for free.
To that end, here is a short, non-comprehensive list of non-monopoly ways to compensate creators (as a theme, they all involve finding something related and _scarce_ to sell):
(1) live performance. Although words and sounds can be easily copied, experiences cannot. Soul Coughing's Mike Doughty discovered this when he found that fans were packing his live performances, singing along with his unreleased songs (that they'd found on Napster).
(See also Wilco.)
Even if they couldn't copyright their words, authors could similarly charge for readings, signings, Q&A sessions, and so on.
(2) patronage. This worked before copyright. It could work again. Pat Dinizio, the frontman for the Smithereens, sold 'subscriptions' to himself which entailed advance music, private concerts, etc...
(3) pay as you go. Steven King tried this, and it didn't work, but I think it was a mistake for him to set a "75% payment" threshhold. If instead he'd set specific money targets, it probably would have worked better. I think we'll see more of this. (See also the Street Performer protocol.)
(4) work for hire. Much software development already takes place this way. Some open source software is written for bounties. And it's easy to imagine someone wanting a song or a book or a short film to give as a gift or in someone's honor.
(5) advertising. Rap songs and movies already include lots of product placements. Some people might find this one distasteful.
And so on and so forth. Some of us _are_ thinking about this.
None of those options are really compatible with public expectations about art in the modern era. They also do not address the core issue Randall brings up, which is the theft of technological advances. We aren't talking about b-sides, but medicines and microchips. A society that OK’s the theft of art with eventually ok the theft of other intellectual property as well. Selective enforcement will lead to no IP rights. No IP rights means no reason to fund major research. No funding, no advances.
Plus all those suggestion you have made essential create a new class division. Those who can afford 'subscription' art and those who can't. Many artists, authors in particular, cannot afford a pay-as-you-go system... nor would their customers readily accept one. THat being said, I do believe for the entertainment business a pay-per-use system is the only feasible solution.
None of those options are really compatible with public expectations about art in the modern era.
Perhaps. But most of the proposed solutions to "piracy" are incompatible with public expectations about personal freedom in the modern era. So something has to give.
No IP rights means no reason to fund major research.
I don't agree with this. There are lots of other reasons to fund major research. (Advantage of being first to market, desire to cure a specific disease, fame, etc...) The question of _whether_ major research would be funded without IP rights is, unfortunately, harder to answer, and I don't think either of us is going to convince the other.
Live performance has been available as an outlet for a long time. It will not replace royalties. Ditto for patronage. Ditto for pay as you go. They are all income sources in addition to royalties.
Yes, some intellectual work will get done. But fewer people will pursue a number of lines of work if the total amount of revenue realizable per worker is less. For instance, some recording artists can suupport themselves from the total income they got from recordings and performances but not if they get revenue from performances alone. On the margin inevitably some will be discouraged from pursuing such a line of work.
Advertising in movies provides a small fraction of the revenue that is derived from movie making. If only advertisements were available many fewer movies would be made.
You say you are thinking about this. Well, I'm not hearing anything new and I'm certainly not hearing any mechanisms that will replace what is being lost as more stuff gets copied in violation of assorted IP laws.
Yes, there are other reasons people fund research. But while NIH spends $28.4 billion the big pharmas fund even more (about $33 bil last I read). Take away incentives to fund research for profit and the total amount of research will go down. Also, the amount of research aimed at producing treatments will go down even more.
To me your argument sounds like an argument made by communists on why even if land property laws were done away with we could still find reasons to get people plow fields, plant and harvest crops, and build buildings. Yes, true enough. But so glad I don't live in the old Soviet Union.
Frankly, from the music artists' PoV, patronage and direct-to-market offers (even of materials that have been bootlegged) are vastly more profitable than the relationship with external copyright holders that there's little reason to hook up with the RIAA for distribution. A number of new artists aren't going in that direction. (This is one of the reasons they've been so uppity about distribution technologies lately; they see their money mine drying up as they get routed around.)
Movies are harder to hook up in that architecture as they require more infrastructure, but they aren't impossible to. Ultimately, the biggest change has ALREADY occured, in that the market now has another means of acquiring the content in competition with the MPAA's. Expectations are now vastly different. Does that mean the day of the movie is over? Of course not. It does, however, mean that the original producers need to find new ways of approaching their markets to avoid being dragged down by the MPAA's floundering. Luckily, indie films have been doing this for a decade or more as the means of film production dropped in price. Pick up a copy of _Rue Morgue_ and flip through, looking for indie films. Can they be bootlegged? Sure. Is the market that's interested in the media interested in FUNDING a bootlegger, or are they more interested in funding the creators? Overwhelmingly the latter.
So, let's map this back to BigPharm(tm). Why do as much creative research in pharmaceutical production if it can be brought to production faster/better by a Chinese company? Because there's a need? (Yes, I realize the idea of researchers doing work to solve problems is hopelessly naive.) Because the Chinese company provides a kick-back in order to make sure the research keeps flowing? But this begs the question ... why should your BigPharm(tm) company be able to profit on something they CAN'T commercialize? ChinesePharm(tm) beats you to market? Poach their best researchers and bring your next product up first. Buy them out wholesale. Or simply double your research efforts to swamp their production capacity; if they grab one, you go with another.
I have this odd faith in capitalistic markets. Its odd, probably brought on by ongoing study of evolutionary forces on living populations. There has been too little pressure for far too long on IP "herding" entities. The pressure is going up. I'm no more able to completely predict what adaptations are going to emerge than you are, but I am perfectly happy to enjoy the fact that pressures are escalating making radical mutations likely to emerge and surprise us. The continuation of the system as-it-was is no longer possible, and we're not in a new stable environment. (We may never be in a new stable configuration, and that's just fine, too.) Like any ecological researcher, there are times we're reduced to simply watching, recording, and waiting.
you want to be able to rip off IP with a clear conscience?
That's completely uncalled for. I happen to be extremely scrupulous in my dealings with intellectual property. I never
make illegal copies of anything.
I also think that the law of intellectual property needs an overhaul and should be dealt with rationally on a strict cost-benefit basis. What is needed is more actual data and less theorizing.
Intellectual property rights are strictly legal rights i.e. they are benefits granted by the government as a means for sponsoring the creation of more ideas, works of art, and inventions. Is that what's happening? The actual data is by no means unambiguous. For example, following the Copyright Extension Act of 1999 applications for copyrights fell.
Some kinds of patents, too, are problematic. I happen to think that software patents and business process patents are absurd. I can point to dozens if not hundreds if not thousands of software patents that fail both on novelty and non-obviousness. I have no explanation for how these patents were granted.
Some patents on the other hand e.g. pharmaceutical patents should probably have longer terms.
You also might be interested in my recent post that explores how government subsidizing of certain kinds of good i.e. intellectual property may be in part responsible for our trade deficit with China.
To repeat what you said:
“Stealability” sounds so immoral. How about “ease of sampling”?
Look, unless a moral argument is made for IP then I do not see how the public can be convinced to respect IP. Utilitarian arguments for why we should all voluntarily sacrifice and not take what we can take are not going to work.
Also, I fully support the moral argument that other people shouldn't be able to take the results of my labor and use it without paying me for it.
More data and less theorizing: Look, we can't run experiments in parallel universes to see how different intelectual property schemes would work in practice. I don't see how we can do much better than we do now in trying to guess which IP schemes would work better. Maybe some economists have come up with some neat tricks for better measuring how well different IP protection schemes would work.
However, we face the non-trivial problem that even if economists could come up with a way to measure which IP protection schemes would produce the most wealth we'd still be faced with people both at home and abroad who are opposed to IP protection and the growth of technology that makes it increasingly easy to rip off IP.
I read your post on China and IP. I think we will keep getting ripped off.
What do you mean by IP rights being "strictly legal rights"? How is that distinct from, say, land ownership rights?
The biggest difference between land ownership rights and IP ownership rights that I can see is that it is a lot easier to detect and stop someone who is violating your land ownership rights.
The problem at hand is simple to state - non-enforceability of property rights. This problem is growing as property is increasingly being defined as information that can be dissascoiated from a physical medium.
Classical economic has trouble dealing with information goods on a couple of fronts. At the most basic level classical economic theory posits that wtillingness to pay equals the marginal cost. With infomration goods the marginal cost is nearly zero suggesting that the price should be zero. Another economic line of reasoning holds that one in valuing an information good one must assess it for its quality and utility. Upon evaluating the infomration the value falls to zero as it is already known.
Yes, copyright law as well as patent law and even plain vanilla property law provides the owner "monopoly" power. The writer misses the point and mistakenly applies the concpet of monopoly to the more basic concpet of property rights. Property right is a core of economic element. One can trade in a market economies that which one owns.