I put things here so they are on the internet

MichaelWeinberg.org

I put things here so they are on the internet

This whitepaper examines the relationship between copyright and 3D printing, focusing specifically on how copyright could be used to hinder the development and growth of 3D printing.

A PDF of this paper can be found here.

A Kindle version of this paper can be found here.

One of the most common questions I get about 3D printing from reporters is “what is going to happen when companies discover this and start freaking out?”  More specifically, what happens when an industry has a “Napster moment” and decides that these crazy 3D printers are destroying their business model?

My usual answer is that I hope that the industry (whatever it is) learns from history and does not simply repeat it.  When the internet became widely available and Napser showed just how easy it was to distribute music online, the music industry’s first reaction was to sue.  The industry dedicated time, money, and effort to trying to turn back innovation and protect a business model that was suddenly outdated.  It took years, and essentially being tricked by Apple, before the industry started dedicating some of that time, money, and effort to monetizing the internet. 

Hopefully, the first industry disrupted by 3D printing skips phase one (sue everyone and pretend the technology does not exist) and jumps right to phase two (figure out how to turn this new technology into a way to make money).  Encouragingly, there is some evidence that this is actually happening.

The first example of this was Teenage Engineering, a company that makes synthesizers.  These synthesizers have small plastic parts, and sometimes those parts break.  Unfortunately, it was costing their customers a lot of money to replace these small plastic parts because shipping them was expensive.  Teenage Engineering decided to put the CAD files for the parts up on its site so that its customers could just print their own.

Today, we have a similar example of an even larger company viewing 3D printing as an opportunity, not a threat.  Nokia announced that it was releasing CAD files for the back shells of one of its phones.  This allows consumers to create their own customized back plates for their phones and print them out on their own.

Simply put, this is great news.  When confronted with widespread access to 3D printing, both of these companies are asking “how can I use this to my advantage?” not “how can I sue to stop this?”  Going forward, hopefully more companies decide to learn from history and follow these examples.

The general reaction to the idea of expanding DRM to 3D printing has been, encouragingly, negative.  DRM has completely failed to slow the supply of unauthorized copies of music, movies, and books online.  At the same time it has succeeded in frustrating perfectly legitimate uses of copyrighted content.  There is no reason to think that either of those outcomes would be different if DRM was applied to 3D printing.  However, there could be a way to apply DRM-like techniques to 3D printing in a positive, consumer-friendly way.

Two distinct categories of DRM

Broadly speaking, there are two types of DRM.  The first is DRM that focuses on the needs of sellers.  This is DRM that is generally associated with music, movies, and books, and basically treats all users like they are criminals a click away from committing copyright infringement.  The second is DRM that focuses on the needs of the buyers.  Truth be told, this is not really DRM at all.  Instead, this is a bundle of techniques that allow users to trace the origin of a particular good.  It has more in common with trademark than copyright, and could also be thought of as digital verification technology (DVT).

Pro-seller DRM, the kind that does not work

DRM, as a stand-alone protection for copyright-protected works, is fatally flawed.  It assumes that as long as an average user cannot circumvent copy protection, no unauthorized copies will be made.  However, because the protected works are digital, an average user does not have to be able to circumvent copy protection.  As soon as one not-so-average user breaks the digital lock and creates an unprotected copy, that copy is available to everyone.  Average users do not need to be able to break the lock themselves – they can just find the unprotected copy made by the person who broke it for them.

Because of this flaw, DRM does not actually provide a benefit to creators.  However, it does impose a cost on users.  DRM can cause platform lock-in, making it hard to transfer things like movies or books between devices.  It can also create a barrier to otherwise legitimate uses – a problem that the Copyright Office tries but fails to address every three years.  

The fundamental weakness of pro-seller DRM is that consumers have no real interest in maintaining its integrity.  A legitimate purchaser who only wants to use media within the scope of DRM-allowed parameters never really comes into contact with it.  A legitimate purchaser who wants to use media in a way that is legal but exceeds the DRM parameters sees it as a nuisance (at best).  An illegitimate copier circumvents the DRM herself or simply finds a copy that has already been separated from DRM.  None of these users has an incentive to support the integrity of the DRM scheme.  

Pro-Consumer DRM, a kind that could work

Contrast this traditional role for DRM with a slightly different application – digital verification technology (DVT).  The role of DVT is not to prevent unauthorized copying on behalf of sellers.  Instead, DVT is designed to assure consumers that the file they have will produce the object they want.

While this type of verification could be used in the digital world (and is used in the form ofchecksums), in general it is not necessary.  A digital copy is, by definition, an exact copy.  A copy of a movie downloaded from iTunes can be copied identically tens, hundreds, thousands of times.  For an end user, the original authorized copy will produce exactly the same movie as the thousandth copy.

The transition from digital to physical makes copying less reliable.  The same digital file can produce meaningfully different physical objects when printed by different types of 3D printers. Even two identical 3D printers will produce slightly different physical objects.

Sometimes these differences will not matter to a consumer.  But other times they will.  It is not hard to imagine that someone printing a functional part for an industrial machine would be interested in knowing that the source file is the file that will produce the correct object on the 3D printer being used, not just a reasonable approximation of that file. 

This is where DVT comes in.  Not every consumer will be interested in using DVT for every print.  Someone buying “designer” sunglasses from a table set up on the sidewalk may not be interested in verifying where the sunglasses were really made or that they were designed correctly.  Similarly, someone downloading a file for designer sunglasses from some dark corner of the internet may not care about the file’s source. 

However, there are plenty of instances where consumers will care.  After all, consumers care about being able to verify that their medicine really came from a factory that was using active ingredients.  They could also care that someone who knew what they were doing designed the latch holding the hood of their car shut – or the bracket supporting the shelf over their head.  This does not mean that the file the consumer is using is an authorized copy.  Rather, it means that the file for the object is actually a copy of the “real” file – not just a file that has been reversed engineered with an unknown degree of accuracy. 

Every person using a 3D printer may not want, or care, to implement this technology.  But it is likely that many will.   Unlike DRM, this consumer interest means that DVT might actually work.  Instead of assuming the end user is a criminal, DVT enlists the consumer as a willing partner.  

A Useful Path Forward

Work is already being done on this type of technology.  For example, Professors Daniel G. Aliaga and Mikhail J. Atallah at Purdue University have been working on embedding signatures in physical objects.  This information could be ignored by an uninterested consumer, but could be highly valuable to a consumer searching for assurances that the file they are using came from a trusted designer and printer.

Of course, there are also individuals and companies discussing more traditional DRM in the context of 3D printing.  They are free to continue what will likely be a wasted effort, at least unless they begin trying to require all 3D printers to implement it.

Going forward, hopefully innovators will focus on ways to give consumers who want it confidence in the source of their digital files and, ultimately, physical goods.  As 3D printing becomes more prominent and 3D printed goods become more common, quality will be critical to convincing the public that the technology is more than a fad.  Allocating development resources in this way, instead of in a futile attempt to prevent unauthorized copies, will be much more beneficial in the long term.

Internet Service Providers (ISPs) regularly insist that data caps are a legitimate tool to ease congestion on their networks and an effective way to signal value to consumers. But, as we haveargued, data caps do not resolve congestion, are confusing to consumers, and lend themselves to unfair and anticompetitive behavior. 

In light of this disagreement, it is a promising sign that a recent study published by the National Cable & Telecommunications Association (NCTA) and co-authored by Steven S. Wildman, the new Chief Economist of the FCC, moves beyond some of the previous rhetoric and takes a significant step towards focusing the debate on real areas of conflict.  

Unfortunately, it stops short of recognizing a critical distinction in understanding the heart of the disagreement. Let’s take a look:

Congestion is not the Issue

The absence of data hogs in this report is encouraging. 
Image created by Vimeo user Nick Cross.

The most refreshing section of the study is the one that is not there.  There is no meaningful discussion of usage-based pricing as a tool to reduce network congestion or a suggestion that monthly data limits are a reasonable way to impact congestion.  There is also no invocation of the mythical “data hog,” a sinful creature that can only be punished with data caps.  Hopefully, the omission is NCTA’s tacit admission of two things: that cable networks are not congested and, if they become so in the future, monthly caps will do little to address that congestion.

Price Discrimination is the Key

Instead, the report focuses on usage-based pricing as a form of price discrimination.  This type of value-neutral economic discrimination (in contrast to other types of discrimination that are morally, politically, and even economically problematic) appears to be one of the key motivating factors behind the imposition of usage-based pricing.  At its core, price discrimination is about making people who value something more pay more for it and allowing people who value something less pay less for it.  This is the way that most of the economy works and there is no reason that price discrimination per se should not be allowed for broadband pricing.

Both types of seats will get you where you are going, but one will cost a lot more. 
Images by flickr user Richard Moross and flickr user johnporcaro.

This is the report’s key argument.  Price discrimination happens every day across our economy.  As an economic practice it creates value generally and can do the same in the world of broadband.  Furthermore, price discrimination can make a form of a good or service available to people who might be priced out without it.

All of this is true and all of this is important to understand before having a meaningful discussion about usage-based pricing and data caps.  However it is the beginning, not the end, of that discussion.  But it is also where the report stops.

Price Discrimination is not the Problem, Data Caps are

The key question about usage-based pricing is not if the general practice of price discrimination has a place in the world of broadband.  After all, ISPs have used speed to implement price discrimination since the beginning of broadband.  Instead, the key question is whether using data caps or usage-based pricing to implement price discrimination has a place in the world of broadband.  On that question the report is silent.

That silence is disappointing.  Public Knowledge’s position is that data caps and usage-based pricing is a type of price discrimination that is especially susceptible to anti-consumer manipulation by ISPs and can suppress activities that we generally encourage.  That is why they warrant attention.

At its core, price discrimination relies on sending signals to consumers.  A consumer values X highly and therefore should pay more for it.  That same consumer sees little value in Y and therefore should pay less.  One of the fundamental problems with using data caps and usage-based pricing as part of a price discrimination strategy is that the signal is hard for consumers to receive.  This deficiency becomes clear when usage-based pricing is compared to the existing price discrimination tool: speed.

Speed is a Signal That is Easy to Understand

Imagine a consumer operating in a world where speed is used to implement price discrimination.  She chooses a speed tier and begins using the internet.  Some things she wants to do work well at the speed she selected.  However, other things do not work as well.  Perhaps pages load too slowly for her liking or video buffers and displays at a low resolution.  Every time a page loads slowly or video buffers she gets a clear signal: “your tier does not support the activity that you are doing at this very moment.”  This signal is temporally connected to the activity and easy to understand – slow load times are because of a slow connection.  

At that moment the consumer can consider how valuable the activity is to her, and how often she has gotten that signal in recent days or weeks.  If the activity is sufficiently important or the buffering occurs with annoying regularity, she may decide to upgrade her tier.  Having used the low tier to its fullest capacity and found it wanting for a collection of known activities, that decision is a reasonably informed one.

Data Caps are Much Harder

Contrast that with a consumer operating in a world with a monthly data cap, the most popular implementation of usage-based pricing.  She selects a cap tier and begins using the internet.  Some things she wants to do work well at the tier she selected.  Others do not.  Unfortunately, it can be hard to tell which is which because she does not find out she has run out of data until her cumulative monthly use exceeds her cap. When she receives that alert, or her bill at the end of the month full of overage charges, she gets a muddled signal: “your tier does not support something, or some combination of things, you did in the past month.” 

At that moment the consumer likely does not remember everything that she did over the past month online, or how it compares to other months.  Did she watch five videos or six?   Was the website she visited two weeks ago featuring abnormally rich interactive content?  Did she upload photos to share with her friends?  Are any, or any combination, of those things worth paying more for?  

In the face of uncertainty, many consumers will come to a frustrating but reasonable conclusion: the best path is to over pay and under use.  As long as they buy a higher tier than they need and avoid trying anything new online, they should be fine.

Focusing on Policy Differences

When Public Knowledge and others raise concerns about usage-based pricing, those concerns are about usage-based pricing, not price discrimination.  Price discrimination that relies on usage-based pricing, especially if it is implemented by an ISP with a pay-video business to protect, raises a number of usage-based pricing concerns. That is why our usage-based pricing whitepaper Know Your Limits ends with a series of usage-based pricing-specific recommendations.  Assertions that price discrimination can help expand access or increase value to consumers may be true, but they are not particularly interesting or relevant to the conversation.

The debate around usage-based pricing can only move forward when people stop talking past each other and start focusing on real policy differences.  By moving past congestion and examining price discrimination, this report marks a significant step towards that goal.  However any argument that focuses on price discrimination alone, or that relies on assertions that are merely true for price discrimination generally, do this issue a disservice.  Hopefully the next report from ISPs will try to explain why usage-based pricing, and not just price discrimination, is a reasonable way forward for the broadband market.

On Friday, Rep. Steve Israel reportedly invoked 3D printing to illustrate his concerns about undetectable plastic firearms.  This may represent the first time that a lawmaker has explicitly linked 3D printing with a perceived threat to society.  Hopefully, Rep. Israel recognizes there is no such thing as a 3D printing-specific solution to the problem of plastic guns.

Let’s be clear about one thing: nobody wants people sneaking guns onto airplanes or into other areas protected by metal detectors.  And we have all seen In the Line of Fire enough times to understand that plastic guns can be smuggled into places that metal guns might never see.  But any attempt to address these concerns should focus on plastic guns, not 3D printers.

Fundamentally, this is because there is not a 3D printing-focused solution to this issue.  3D printers work by turning a digital file into a physical object.  That physical object can be made out of any number of materials.  While a digital file may work better or worse with some materials than others, nothing about the file is inherently tied to a given material. 

This versatility is one of the characteristics that makes 3D printing so powerful.  Even without a 3D printer of your own, you can buy Bathsheba Grossman’s Gyroid sculpture in plastic.  Or steel.  Or silver.  Or glass.  Or sandstone.  You could travel to a Dutch Staples and print it in paper.  You could make nice with Hod Lipson and Jeff Lipton at Cornell and print it in cheese or batter.  That is just as true for a gun file.  The difference between a detectable and undetectable 3D printed gun lies in the printing.  Not the file.

And that file is a digital file just like any other.  For all intents and purposes, once it exists in one place online it exists every place online.  Congress cannot make a file for a 3D printed gun disappear off the internet any more than it can do so for a pirated version of The Avengers.

Finally, there is no way to stop a 3D printer from printing a particular type of thing.  As we havewritten before, a 3D printer is a general purpose machine that can be used to make just about anything – both good anythings and bad anythings. Printers do not run software that checks with a central approved database before they print something out.  In fact, because 3D printerscan print themselves, there is not even a central database that keeps track of all of the 3D printers in the world.

That is why, as a general rule, we do not focus on tools when we are trying to solve a policy problem.  We would never try and stop bombs by passing a law controlling wires or try and stop hacking by passing a law controlling the use of command line tools. 

Instead, we focus on behaviors.  It is illegal to blow something up with a bomb no matter how it was made.  It is illegal to hack into protected database no matter how you do it.   If you believe that people should not be able to carry plastic guns, make it illegal to possess plastic guns.  No one is worried about printing undetectable guns with a 3D printer per se.  They are worried about undetectable guns. 

It is probably a good thing that Rep. Israel is pursuing a conversation about plastic guns, and these days invoking 3D printing is a good way to bring attention to an issue.  But in crafting a solution, hopefully Rep. Israel will focus on the problem – not just a high-profile tool.  After all, if Congress passed a law banning undetectable guns in 1988, and John Malkovitch’s Mitch Leary was creating plastic guns in 1993, eliminating 3D printing from the equation is not going to solve any problems.  But it could cripple the growth of legitimate applications for this promising technology.