DRM as an Enabler of Business Models:

ISPs as Digital Retailers

 

by Lionel S. Sobel *

 

 

1 I.              The quest for better business models

2 II.            DRM-based business models

3                 A.            Anti-copyright model

4                 B.            Beyond copyright models

5                 C.            Copyright based models

6                                 1.             Statutory license models

7                                 2.             Models giving copyright owners discretion and control

8 III.           ISPs as Digital Retailers Model

9                 A.            The role of control

10                 B.            How the Digital Retailer Model would work

11                                 1.             Enabling DRM technologies

12                                 2.             Implementation of technology by ISPs

13                                 3.             Statutory license

14                 C.            The Digital Retailer Model compared to the Tax and Royalty System

15                 D.            Objectives satisfied

16                 E.             Problems requiring solutions

17                                 1.             Spamming

18                                 2.             Intra-industry conflicts

19                                 3.             Privacy

20                                 4.             Pay-per-use, fair use and non-infringing uses

21                                 5.             Unregulated royalty rates

22 IV.           Conclusion

 

23 I.              The quest for better business models

24                 Better business models are the Holy Grail of the digital age. Even Alex Doonesbury has joined the quest.[1] “Alex” is the daughter of “Mike Doonesbury” in Garry Trudeau’s syndicated comic strip, so her search can take as long as Trudeau likes. Things are more urgent for those in the real world, especially for those in the entertainment industry. Digital reproduction and distribution have shattered traditional music industry business models, and are on the verge of doing the same to movie industry models. The question is what, specifically, can be done about it; and it’s a difficult question to answer.

25                 Proof that the question is difficult (if proof be necessary) can be found in the pages of the nation’s leading business periodical, The Wall Street Journal. In an editorial (triggered by a U.S. Naval Academy investigation of midshipmen suspected of downloading MP3 files), the The Wall Street Journal recently advised the record industry, quite unhelpfully, that it “needs a new business model.”[2] The editorial pages of the Journal usually trumpet the interests of big business. But this particular editorial criticized the record industry as one “still wedded to an LP-era business model”; and it was dismissive of “MusicNet” and “pressplay,” the record industry’s maiden tests of new models based on digital distribution.

26                  The Wall Street Journal did not, however, offer particulars for any business model it would find praise-worthy. And therein lays the rub. It’s one thing to say a new model is “necessary.” It’s quite another to suggest how that model might work. The devil – as many have noted – is in the details.

27                 Mindful that seemingly attractive concepts are often undone by their necessary details, I nevertheless suggest a business model that in my view has promise: Internet service providers (ISPs) should become digital retailers. They should license digital works of all kinds – music, movies, television programs, photographs and other graphic images, books and periodicals, and even software – from the owners of the copyrights to those works, at wholesale prices set by copyright owners. And ISPs should sell those works to their subscribers, at retail prices set by ISPs themselves.

28                 The profit potential for ISPs should give them reason to embrace this model. Consumers should embrace it too, because it would give them the choice and convenience they crave, though they’d have to pay for what they buy. Moreover, digital middlemen – like website operators, P2P networks, newsgroup and chat room hosts, Internet search engines, and online radio and television stations – would be able to serve as promoters and distributors, for love or money or both, free from potential copyright liability of any kind (direct, contributory or vicarious). And computer and consumer electronics manufacturers and software companies would be able to invent and innovate to the best of their abilities, without regulation of their products’ designs.

29                 Digital Rights Management technology (DRM) makes this model possible. In a conference devoted to “The Law & Technology of DRM,” it may not be necessary to defend the use of DRM or persuade people of its value. But some particular applications of DRM have been controversial, so two preliminary comments are warranted.

30                 First, though DRM is young, it and its legal status are formally recognized already, worldwide, in at least 39 nations that have become parties to the WIPO Copyright Treaty.[3] The Treaty refers to DRM as “technological measures” used to exercise rights and restrict unauthorized acts, and as “copyright management information” used to identify authors, rights holders and the terms of authorized use. And the Treaty requires adhering nations to provide legal protection for both.[4]

31                 Second, DRM will be at the foundation of whatever business models actually succeed in the digital age. Though The Wall Street Journal didn’t offer the record industry any particulars for a new business model, others have; and at least some DRM features have been at the heart of all of the suggested models.

32 II.            DRM-based business models

33                 Digital business models – those that have been tried already, and those that have been suggested – exhibit a phenomenal breadth of opinion, ranging from what I refer to as “anti-copyright” to “beyond copyright” models. To appreciate the details of a business model that would make ISPs into digital retailers, it will helpful to locate that model along the “anti-copyright” to “beyond copyright” spectrum.

34                 A.            Anti-copyright model

35                 The anti-copyright model would eliminate copyright entirely, in the online digital domain.[5]  DRM may play a role even in this model, but only to identify works’ authors who audiences may choose to compensate with “tips.” In this model, DRM would not be used (or even legal) to prevent unlicensed copying or redistribution of works.

36                 B.            Beyond copyright models

37                 At the other extreme is a model that would go beyond copyright. Publishers would use DRM to control access to works, even those in the public domain, and to prevent unauthorized copying and redistribution of those works. Access would be controlled by passwords, and unauthorized uses would be prevented by encryption and watermarks. Circumvention would be banned and punished by law. The public domain status of these works would be recognized, by allowing others to digitize and distribute their own versions of those works, at their own expense, without liability. But this model would prevent others from copying existing digital versions – would prevent them, in other words, from taking a free ride on the investments made by other companies that previously digitized works, even those in the public domain.

38                 Another business model also goes beyond copyright, but not quite as far. It is a model that uses DRM to control access to public domain materials, but doesn’t control copying or redistribution of those works. In this model, digital versions of public domain works are distributed to subscribers, unencrypted; and copying and redistribution is controlled, if at all, merely by contract. This model is used by Westlaw and Lexis, and even by the United States government in its pay-per-page PACER system for the digital distribution of federal court judicial decisions.[6]

39                 C.            Copyright based models

40                 Between the “anti-copyright” and “beyond copyright” models are two batches of models, both of which recognize and respect copyright. One batch would impose statutory licenses that authorize digital uses of copyrighted works. The other would give copyright owners discretion over licensing terms, and control over unauthorized uses of their works.

41                                 1.             Statutory license models

42                 One statutory license model is Neil Netanel’s proposed “Noncommercial Use Levy.”[7] It would permit noncommercial copying, distribution, performance and even adaptation of copyrighted works, in return for levies collected from providers of those products and services whose value is enhanced by file swapping. Collected levies would be allocated among copyright owners by category (record companies, movie producers, book publishers, and so forth), and then among individual copyright owners within each category. The amount of the levy would vary among products and services, and would be determined in Copyright Office arbitrations (unless affected industry segments agreed on levies themselves). Allocations among those entitled to receive levies would be in proportion to the extent to which their works were used. Disputes about allocations presumably would be determined by Copyright Office arbitrations as well.

43                 A second statutory license model called the “Tax and Royalty System” has been proposed by Terry Fisher (apparently with more enthusiasm lately than at first).[8] Under this system, a “tax” would be assessed on ISP access and on technologies used to perform music, including MP3 players, hard drives and even computers; and the revenues from these assessments would be distributed to copyright owners in proportion to which their works are accessed. (Professor Fisher focuses on the recorded music industry in particular; but there is no reason his Tax and Royalty System couldn’t be used to compensate copyright owners in other industries too.)

44                 The “Noncommercial Use Levy” and the “Tax and Royalty System” both use DRM to determine the extent to which particular copyrighted works are used. Digital copies of works would be embedded with watermarks; ISPs would detect and record those watermarks as files flow through their servers; lists of detected works would be compiled periodically, along with the frequency with which they were detected; and that data would be used to allocate collections proportionately among copyright owners.

45                 The two key features of both the “Noncommercial Use Levy” and the “Tax and Royalty System” – royalty setting and royalty allocation – are based on well-established elements of existing copyright law. Statutory license fees already are set, and collected fees allocated, by Copyright Office arbitrations in connection with two types of uses of four kinds of works: cable and satellite retransmissions of copyrighted movies and television programs, and the musical compositions in their soundtracks,[9] and consumer duplication of digital music recordings.[10] The license fees for certain online digital performances of music recordings also are determined by Copyright Office arbitration[11] (though allocations of digital performance royalties among those entitled to receive them was done by Congress, in the Copyright Act itself, rather than by the Copyright Office[12]).

46                 Although the “Noncommercial Use Levy” and the “Tax and Royalty System” are similar, they differ in at least one important respect. The “Noncommercial Use Levy” would permit users to create new versions of digital works, in addition to making and redistributing copies. The “Tax and Royalty System” does not seem to contemplate the creation of new versions; it would simply authorize copying and redistribution. This means the “Tax and Royalty System” protects copyrights somewhat more than the “Noncommercial Use Levy,” because the “Tax and Royalty System” leaves more control in the hands of copyright owners, namely, the right to license the creation of new versions of their works, on terms agreed to in private negotiations.

47                                 2.             Models giving copyright owners discretion and control

48                 Several models give copyright owners discretion over licensing terms and control over unauthorized uses of their works.

49                 One model gives copyright owners access control, using passwords as the only DRM feature. Content is not encrypted, and this model does not control copying or redistribution. Familiar examples of this model include the online editions of the New York Times and Wall Street Journal. Both websites require registration to obtain a password that is necessary for access. Use of the Times site is free to anyone. The Journal’s site, by contrast, requires payment of a subscription fee (even by those who subscribe to the regular, paper edition). Both companies enforce their password requirements using technologies that run on their own servers. Users’ computers do not require password-related design features or website-specific software.

50                 A second model gives copyright owners access control, plus copy and redistribution control too. Control is accomplished with DRM, namely, encryption that both restricts access and controls what may be done with materials by those who are given access. Special software (provided by or on behalf of copyright owners) is necessary to get access to the encrypted material. And that same software enables authorized uses of it – but not unauthorized uses – by those who are entitled to access. Familiar examples of this model include publications in the Adobe eBook format, and audio and video materials in the RealMedia and Windows Media formats.

51                 A third model gives copyright owners control over access, but not over copying or redistribution, using encryption. This model requires authorized users to have specially designed equipment to receive and decrypt materials. But companies that manufacture the necessary equipment voluntarily incorporate the necessary design features into their equipment; they are not compelled to do so by law. This model is used by cable systems and satellite TV companies.

52                 A fourth model gives copyright owners control over access and over copying and redistribution, using encryption. This model too requires authorized users to have specially designed equipment to access and decrypt materials. Manufacturers of this equipment incorporate necessary design features voluntarily; they are not required by law to do so. This model is used in connection with movie DVDs which are encrypted and then decrypted using the Content Scramble System (commonly referred to as “CSS”).[13] The record industry’s Secure Digital Music Initiative would have used this same model.[14] And though SDMI (as the initiative was commonly known) wasn’t implemented in connection with commercially-released CDs, similar technologies now being used by some record companies are based on the same model.[15]

53                 Finally, a fifth model uses DRM to give copyright owners access, copy and redistribution controls over digital works that are not encrypted, but are watermarked with authorized-use information. Because these works are not encrypted, this model works only if computers and consumer electronics devices contain circuitry that recognizes and responds to watermarks. (Without such circuitry, computers and other devices would simply play unencrypted works, and permit them to be copied and redistributed.) One example of this model is the Serial Copy Management System, intended to permit record companies to control digital copying of recorded music. This System is at the heart of the Audio Home Recording Act of 1992 which requires digital audio recorders to be equipped with circuitry that prevents them from being used to make serial copies (that is, copies of copies) of digital recordings.[16] The commercial significance of that ban was largely undercut by the advent of MP3 technology for storing recorded music, and consumers’ use of computers, rather than digital audio recorders, to copy and redistribute MP3 files. MP3 technology undercut the Serial Copy Management System, because the Audio Home Recording Act exempts computers from the need to have anti-copying circuitry.[17]

54                 Nevertheless, this fifth model remains at the forefront of current debates, because unencrypted digital television broadcasting is on the near horizon. The FCC has mandated the introduction of digital television broadcasting, nationwide, by 2006.[18] But movie and television producers aren’t going to provide expensive content for digital TV broadcasts, if that content can easily be copied and forwarded over the Internet to recipients around the world. As a result, the lack of effective copy protection methods may hinder the development of digital TV broadcasting by greatly reducing the amount of attractive programming that is made available for it.

55                 The copy protection method proposed for digital television broadcasts is called the “Broadcast Flag System.” It was the centerpiece of a bill in the 107th Congress formally entitled the “Consumer Broadband and Digital Television Promotion Bill.”[19] The bill – commonly referred to as the “Hollings Bill” – would have required “digital media devices” to provide “effective security for copyrighted works.” The 107th Congress adjourned without enacting – or even voting on – the Hollings Bill. But that doesn’t delay the effective date of nationwide digital TV broadcasting.

56                 With these developments in mind, the FCC recently issued a Notice of Proposed Rulemaking by which the Commission invited comments on whether it should adopt rules that would mandate the incorporation of copy protection technology into television receivers and other consumer electronics devices, such as digital TV recorders.[20] An alliance of copyright owners, broadcasters, and entertainment industry unions has urged the FCC to adopt a rule that would require devices to recognize and respond to “Broadcast Flags” included in digital TV broadcasts – Flags that would indicate whether those broadcasts may be redistributed outside the recipient’s home.[21] Broadcast Flags do not encrypt digital TV signals, and those signals will be broadcast unencrypted. So the Broadcast Flag System will work only if devices that receive and process digital broadcasts are designed to recognize whether particular signals may be redistributed outside the recipient’s home, and only if those devices do not permit redistribution if a signal’s Broadcast Flag does not authorize it.

57                 The following chart recaps these business models (excluding the anti-copyright and beyond copyright extremes). Those that protect copyright the most are at the top of the chart; those that protect copyright the least are at the bottom. The chart also reflects the technology required to implement each business model. Not coincidently, the chart shows that in order to provide more control over copyright, more control over technology must be provided too – under the business models discussed thus far.

 

58 Copyright Control	59 Business Models	60 Technology Requirements	61 Control over Technology
62 More                         63 Less	64 Access plus copy and redistribution control over copyrighted materials (SCMS for digital audio music recorders; Broadcast Flag System for digital TV broadcasts)	65 Requires equipment containing legally mandated features	66 More                         67 Less
	68 Access plus copy and redistribution control over copyrighted materials (CSS for DVDs; SDMI for music CDs)	69 Requires special equipment with voluntarily installed features
	70 Access but not copy or redistribution control over copyrighted materials (Cable and satellite TV)	71 Requires special equipment with voluntarily installed features
	72 Access plus copy and redistribution control over copyrighted materials (Adobe eBook, RealMedia, Windows Media)	73 Requires only software provided by or on behalf of copyright owner
	74 Access but not copy or redistribution control over copyrighted materials (New York Times/Wall Street Journal online)	75 No special technology required on users’ equipment
	76 Tax and Royalty System	77 No special technology required on users’ equipment
	78 Noncommercial Use Levy	79 No special technology required on users’ equipment

 

80 III.           ISPs as Digital Retailers Model

81                 A.            The role of control

82                 The ISP as Digital Retailer model provides more control over copyright with less control over technology than any one of the models just described. This balance should make the Digital Retailer model more attractive than others, because copyright owners want to control their copyrights while technology companies want to create their products without legally mandated controls.

83                 Controls over copyright and technology are not ends in themselves, even for copyright owners. Instead, control over whether works can be copied and redistributed is important to copyright owners, because unauthorized copying and redistribution destroys their ability to market their works, in two ways.

84                 First, copyright industry business models are based on the strategy of selling multiple copies of works, or performing them multiple times, at per copy or per performance prices that are just a tiny fraction of the cost of producing those works. Uncontrolled copying and redistribution destroys this plan, because unauthorized digital copies displace those sales and performances.

85                 Second, unauthorized copying and redistribution of copyrighted works prevents copyright owners from pricing their works in ways they hope will maximize their incomes.

86                 Pricing strategies are at the core of the “marketing” portion of every thorough business plan. Ideally, sellers – including copyright owners – would charge higher prices to those customers who value the sellers’ goods or services most and are best able to afford higher prices, and would charge lower prices to customers who value the sellers’ goods or services less or are less able to afford them. Economists refer to this as “price discrimination,” but it’s not “discrimination” in the civil rights sense of the word. In business, price discrimination is a good thing, because so long as sellers receive more than their marginal costs, additional sales are profitable. Profits can be maximized in this fashion, if but only if sellers aren’t required to reduce the prices they charge those who value their goods and services the most, simply because sellers charge lower prices to other less motivated or wealthy customers.[22]

87                 It’s been argued that copyright owners do not have the ability to engage in perfect price discrimination.[23] The same of course could be said of sellers in every industry. (Even business travelers book plane reservations in advance, or stay over a Saturday night, when they can, in order to get lower fares intended for vacation travelers.) But successful business plans do not require perfect price discrimination. They simply require the ability to price discriminate a little, as in these familiar examples:

88 ·         Hardcover books are published before, and cost more than, paperback reprints.

89 ·         Movies are exhibited in theaters before they are available on DVDs, and movie theater tickets cost more than DVD rentals (which cost more than viewing movies on pay-TV, which cost more than watching them on advertiser supported TV).

90 ·         It costs more to buy a DVD or videocassette than to rent it.

91 ·         New albums by musical artists cost more than “greatest hits” compilations, which cost more than multi-artist albums compiled by theme.

92 ·         Full-featured versions of computer software cost more than “lite” versions, which cost more than “trial” versions.

93                 The point is that business plans for the marketing of copyrighted works are based on the ability to do sequential but separate releases of those works. And in entertainment businesses, the sequence for successful works is spread over a long time. Uncontrolled copying and redistribution of works interferes with this sequential release.

94                 Technology companies have similar concerns, but for them, design innovations are central to their business plans. They fear that legal regulation of their products’ features will interfere with product innovation, and thus with their business plans.

95                 B.            How the Digital Retailer Model would work

96                                 1.             Enabling DRM technologies

97                 In order for ISPs to become digital retailers, digital versions of copyrighted works have to be identified, digitally, so their purchase can be tracked electronically. This can be done, using two types of existing DRM technologies: “watermarking” and “fingerprinting.”

98                 Watermarks are digital identifications inserted into digital copies of works at the time they are manufactured.

99                 Not all digital works will have watermarks. Legacy (that is, older) works (including older digital works) were created without watermarks. Digital copies of analog works (such as music cassettes and video tapes, photographs, and texts) do not have watermarks. Even digital copies of works that were watermarked originally may have been stripped of their watermarks when they were converted to analog copies and then redigitized.

100                 Copyright owners can create digital identifiers for unwatermarked copies of their works by “fingerprinting” them. Fingerprinting converts the work’s own content into a unique digital identification mark, by applying an algorithm (or mathematical formula) to selected features of that content.[24]

101                  Together, watermarking and fingerprinting can be used to create digital identifications for every digital work that copyright owners want to have identified. These identifications then can be used to recognize works transmitted online from websites, over P2P networks, as email or instant message attachments, and in any other way that involves digital files moving through networks to which users connect through ISPs.

102                                 2.             Implementation of the technology by ISPs

103                 Under the Digital Retailer model, the technology to used to identify watermarked and fingerprinted files would reside on ISPs’ servers, not on end-users’ computers or consumer electronic devices. This is key, for three reasons.

104                 First, putting the technology on ISPs’ servers frees technology companies to innovate at will, without legal regulation of their products’ designs.

105                 Second, putting the technology on ISPs’ servers makes circumvention less likely. (CSS, Adobe eBook and SDMI – all of which are implemented on consumers’ computers – were circumvented quite quickly.[25] And one recent technical report persuasively argues that watermark detection technology on users’ computers or electronic devices could be easily defeated, if implemented in software, and would make computers and devices obsolete too quickly, if implemented in hardware.[26])

106                 Third, all online service users connect to the Internet through ISPs, and thus can be billed by their ISPs for whatever copyrighted works they access. (ISPs already are able to meter the bandwidth usage of each of their subscribers. And it was announced last year that ISPs may begin charging subscribers based on usage, rather than flat monthly fees.[27])

107                 ISPs would monitor the flow of copyrighted works through their servers, looking for watermarks and recording the recipients of watermarked files. A database would identify the owner of the copyright to each watermarked file, as well as the wholesale price the copyright owner decided to charge for its use.

108                 Files that were not watermarked would be checked against a fingerprint database, which (like the watermark database) would identify the owner of the copyright to each of those files, as well as the wholesale royalty the copyright owner decided to charge for its use. In order for works to appear in a fingerprint database, copyright owners would have to arrange for their works to be fingerprinted and included in the database.

109                                 3.             Statutory license

110                 Copyright owners would be obligated, by statute, to permit the copying and redistribution of their works. But they wouldn’t be obligated to watermark or fingerprint their works. If they didn’t, those works would be free, to ISPs and to their subscribers. On the other hand, for watermarked and fingerprinted works, ISPs would be obligated, by statute, to pay the royalty charged by each work’s copyright owner. This proposal amounts to a statutory license (because it authorizes copying and redistribution of copyrighted works, without negotiated licenses from copyright owners). But it’s a two-edged statutory license: it authorizes the use of copyrighted works, but also requires ISPs to pay royalties at whatever rates are set by copyright owners.

111                 ISPs would not have to bear the cost of copyright royalties, themselves. Instead, they would be authorized to charge subscribers for watermarked and fingerprinted files they receive, at whatever prices ISPs choose to charge. In most cases, I envision a markup of 100% or so. That is, about 50% of the retail price paid by subscribers would be retained by ISPs, and about 50% would be paid to copyright owners. This is the traditional split between retailers and publishers in the book business, between retailers and record companies in the music business, and between theater owners and distributors in the movie business. However, some ISPs may use lower retail prices as a competitive tool, to attract subscribers from ISPs that charge higher prices. And some ISPs may offer subscribers bulk-purchase plans (just as cell phone companies offer local and even long-distance packages as an alternative to minute-by-minute charges).

112                 C.            The Digital Retailer Model compared to Tax and Royalty System

113                 If the Digital Retailer Model were to be placed in the business model chart, it would be slotted above the Tax and Royalty System, but beneath the others, thus:

 

114 Copyright Control	115 Business Models	116 Technology Requirements	117 Control over Technology
118 More 119                                120 Less	121 Access plus copy and redistribution control over copyrighted materials (SCMS for digital audio music recorders; Broadcast Flag System for digital TV broadcasts)	122 Requires equipment containing legally mandated features	123 More 124                                125 Less
	126 Access plus copy and redistribution control over copyrighted materials (CSS for DVDs; SDMI for music CDs)	127 Requires special equipment with voluntarily installed features
	128 Access but not copy or redistribution control over copyrighted materials (Cable and satellite TV)	129 Requires special equipment with voluntarily installed features
	130 Access plus copy and redistribution control over copyrighted materials (Adobe eBook, RealMedia, Windows Media)	131 Requires only software provided by or on behalf of copyright owner
	132 Access but not copy or redistribution control over copyrighted materials (New York Times/Wall Street Journal online)	133 No special technology required on users’ equipment
	134 ISPs as Digital Retailers	135 No special technology required on users’ equipment; necessary technology implemented on ISPs’ servers
	136 Tax and Royalty System	137 No special technology required on users’ equipment
	138 Noncommercial Use Levy	139 No special technology required on users’ equipment

 

140                 The Digital Retailer Model is similar to the Tax and Royalty System, because both rely on ISPs to collect royalties from their subscribers, both use DRM to identify digital works accessed over the Internet, and both use DRM-enabled identifications to allocate collections among copyright owners entitled to receive royalties. But in my view, the ISP as Digital Retailer model is preferable to the Tax and Royalty System.

141                 The Digital Retailer model gives technology companies just as much freedom to innovate with new product designs as the Tax and Royalty System does, while the Tax and Royalty System deprives copyright owners of the freedom to innovate with new pricing plans at least as much as the Broadcast Flag System deprives technology companies of the freedom to innovate with new product designs.

142                 The Tax and Royalty System requires expensive and time consuming legal proceedings, both to establish royalty rates and to distribute collected royalties. Simply determining the statutory license fee for digital transmission of recorded music has required: an arbitration, a decision by the Librarian of Congress, a Copyright Act amendment, a privately negotiated interim agreement, and a still-pending judicial appeal.[28] In a separate matter, a Copyright Office arbitration costing more than $41,000 was necessary to resolve conflicting claims to digital audio recording royalties totaling just $6.10, split between two songwriters and music publisher.[29] Moreover, each of these proceedings involved works of just one type: recorded music (in the digital transmission proceeding); and musical compositions (in the digital audio recording proceeding). Under the Tax and Royalty System, rate setting and royalty distribution proceedings will be infinitely more complex than in those proceedings, because digital works range from $600 computer programs (like Photoshop) to $1 recorded music tracks. Leaving software out of the plan altogether doesn’t solve the problem, because even though unlicensed MP3 files have been the most newsworthy, the problem includes unlicensed redistribution of computer programs too.

143                 I acknowledge that tracking copyrighted works would be less cumbersome under the Tax and Royalty System than under the Digital Retailer model. The Tax and Royalty System may be able to allocate collections using data obtained by digital file sampling (the way ASCAP and BMI sample radio play of musical compositions, in order to allocate public performance royalties). The Digital Retailer model, by contrast, contemplates complete file tracking and end-user billings.

144                 Nonetheless, in personal conversations with technology vendors, I have been advised that technology already exists that will track individual watermarked and fingerprinted files, without slowing network activity. (I also acknowledge that when I repeated this information to a movie industry executive, she responded that the key word, in what I told her, was “vendor.”) Even if existing technology cannot yet handle the job it would have to perform in order for the Digital Retailer model to work, I believe the issue would be one of scale rather than function. That is, existing technologies may have to be improved to handle the large-scale task, but it doesn’t appear that any new technologies would have to be invented.

145                 D.            Objectives satisfied

146                 Having ISPs serve as digital retailers achieves several objectives.

147                 For copyright owners, it is a model that enables them to be paid for all uses of their works – downloads, streams and attachments – at royalty rates they set themselves.

148                 For consumers, it is a model that gives them ready, legal access to digital versions of copyrighted works, though they would have to pay for what they receive (just as we do in the physical world).

149                 For website operators, P2P networks and users, emailers, instant messengers, online indexes and search engines, it is a model that enables them to play whatever role they desire in the distribution of digital works, legally, without need to get any further consent of copyright owners. They may do this for love, or even for money. Online indexes and search engines would be free to charge for their use, or sell advertising space on their display pages; and they could do so without sharing their revenues with copyright owners. (Copyright owners would be paid by ISPs, if and when works were accessed.)

150                 For computer and consumer electronics manufacturers and for software companies, it is a model that allows them to build and sell their products, without any legal constraints on how they are designed, and without any legal requirement that they contain, or not contain, certain features.

151                 For ISPs, it is a model that gives potential customers incentives to subscribe to broadband service, and it gives ISPs a significant additional revenue source – one that is likely to be equal in size to the revenues received by copyright owners from the online distribution of copyrighted works.

152                 E.             Problems requiring solutions

153                 In order to implement the Digital Retailer model, I acknowledge that a number of problems would have to be solved.

154                                 1.             Spamming

155                 Since virtually all works transmitted online are eligible for copyright protection, and all copyright owners would be entitled to be paid at rates they set themselves, unscrupulous authors may attempt to “game” the system, by spamming recipients with unwanted material, in order to get royalties. Technology may provide a solution to this problem, but if not, another solution is available.

156                 ISPs’ servers will be alerted to the existence of copyrighted material by watermarks or fingerprints, at or before the moment those files are transmitted to Internet users. As a matter of technology, it would be possible for ISPs to send pop-up notices, informing users that files requiring payments are about to be sent, along with their cost, before the actual files are transmitted. Users would then be given an opportunity to click an on-screen button, indicating whether or not they want the files sent. To users, the process would look exactly the way virus warnings look today; and users would respond, the way they respond to virus warnings, with a simple click of the mouse.

157                 The other, non-technical, solution is drawn from the world of credit card fraud. In order to receive copyright royalties under the Digital Retailer model, identification information for materials sent by spammers would have to be placed in watermark and fingerprint databases (along with the watermarks and fingerprints of other copyright owners). ISPs could be authorized to suspend royalty payments to those against whom spamming complaints are lodged, just the way banks suspend or revoke the credit card merchant accounts of retailers if consumer complaints are lodged against them.

158                                 2.             Intra-industry conflicts

159                 To implement the Digital Retailer model, conflicts within the entertainment industry would have to be resolved. Two such conflicts come immediately to mind.

160                 The first is the result of an old but still troublesome fact; single works often embody several separately owned copyrights.

161                 Music recordings embody at least two copyrights per track: a copyright in the musical composition, usually owned by a music publishing company (though if a song is co-written by more than one songwriter, the musical composition copyright is likely to be co-owned by more than one publisher); and a copyright in the recording itself, usually owned by a record company. As a result, royalties for the online performance or download of a single recording must be split between two (or more) copyright owners. What’s worse, music publishers license performances and downloads through separate agencies (ASCAP, BMI or SESAC for performances; and the Harry Fox Agency for downloads). So today, royalties for the online use of a single music recording may be claimed by three separate agencies on behalf of two (or more) separate copyright owners.

162                 Movies too may embody several separate copyrights: one in its visual elements and the sound effects in its soundtrack; and another in each song in the soundtrack. As a result, royalties for the online performance or download of a single movie may have to be split among two or more copyright owners.

163                 Some copyright owners may demand too much, and thereby discourage customers from making online uses of works to which those copyright owners contributed. Other contributors to the same work may be pressured to decrease their royalty rates, in order to lower the total royalty claimed for that work enough to prompt sales of it. This, however, could trigger strategic bargaining among copyright owners – each hoping to persuade the others to lower their royalty demands – a process that may not succeed in lowering the total royalty enough to actually increase sales.

164                 Under the Digital Retailer model, none of these kinds of conflicts is of concern to ISPs. But before ISPs can know who to pay, conflicts like these will have to be resolved.

165                 The second intra-industry conflict could affect ISPs. Some ISPs and some copyright owners are subsidiaries of the same corporate conglomerate. Given complete discretion, such a conglomerate may choose to implement a business plan that seeks to attract subscribers to its ISP subsidiary by offering them exclusive access to the conglomerate’s copyrighted works, or access at lower rates than charged to unaffiliated ISPs.[30] The Digital Retailer model, however, would give all ISPs access to all copyrighted works. And to prevent copyright owners from substituting high prices for exclusivity, all ISPs would have to be charged the same wholesale royalty for each work. Copyright owners could not favor some ISPs with lower royalties than they charge other ISPs. This means, for example, that Warner Bros. and Time could not give charge their sister company AOL lower royalties for digital recordings or online magazines than they charge other ISPs, let alone give AOL an exclusive.

166                                 3.             Privacy

167                 The Digital Retailer model requires copyright owners to make significant concessions, over who, how, when and where their works are distributed. And it requires concessions from users too, most obviously in the area of privacy. In addition to paying for copyrighted works, users will have to tolerate some loss of privacy.

168                 The Digital Retailer model requires ISPs to compile records of copyrighted works accessed by their subscribers, for billing purposes. I characterize this as a loss of “some” privacy, because the degree of the loss should not be measured from a baseline of complete privacy. Credit card companies already know where we shop and how much we spend; and when we shop in places, or spend amounts, that look unusual, company employees call us on the phone to ask whether we used our cards, or thieves did. Likewise, phone companies already know who we call and when, and how long we talk. Even cash transactions in brick-and-mortar retail stores are likely to be videotaped. And in many American cities, highway and toll bridge users and their passengers are likely to be videotaped as well.[31]

169                 We actually enjoy very little privacy today. Tracking copyrighted works we access online, for billing purposes, diminishes our privacy very little further.

170                                 4.             Pay-per-use, fair use and non-infringing uses

171                 Finally, some may object to the Digital Retailer model on the grounds that it requires the payment of a pay-per-use royalty, and makes no provision for fair (or other non-infringing) uses without payment. The objection is factually accurate, though only in part, and in any event, is no reason to reject the model.

172                 The Digital Retailer model does not require payment for each use of a work. It requires payment each time a work passes through an ISP’s server. So, rather than characterizing the model as a pay-per-use model, it should be thought of as a pay-per-redistribution model. Downloaded works may be used on the computer to which they are downloaded, countless times without additional payment. Only the initial download triggers a royalty fee.

173                 The fair (and other non-infringing) use objection implies that users should be able to get access, for free, to copyrighted works they intend to use in ways that qualify as non-infringing. That, however, has never been the case in the physical world. Teachers, for example, may be entitled to display or even photocopy newspaper articles for use in their classes; but that does not mean they have the right to take, for free, copies of newspapers from the newsstands they pass on their way to school. Likewise, movie critics have the right to include plot synopses and quote dialogue in their reviews; but they are not entitled to free admission to movie theaters showing the movies they intend to review.

174                                 5.             Unregulated royalty rates

175                 I have left unregulated royalty rates for last, because, for me, they are not a problem at all. I acknowledge, however, that for others, they may be. That is, I imagine that others might argue that the ability to legally reproduce copyrighted works and redistribute them online cannot be taken advantage of, if copyright owners can charge whatever they wish, whenever their works are copied and redistributed online. Copyright owners can use very high royalty rates – it might be argued – as a technique for preventing online copying and redistribution of their works, in actual practice.

176                 While it is true that copyright owners may charge high royalties for some works, especially when they are new, they always have had the ability to do so in the physical world. (A recent report entitled “Digital Rights Management: Content Protection in the Networked Economy” has been priced by its publisher at $995,[32] and newsletters published by the same company cost more than $1,000 a year[33] – without apparent objection from anyone.) There’s no reason things should be different – especially not by law – in the digital world.

177 IV.           Conclusion

178                 In a perfect world, technology companies would be able to design their products as they think best, and copyright owners would be able to market their products as they think best. Digital copying and redistribution have made these objectives incompatible, at least in part. The quest is for a business model that best accommodates these conflicting objectives.

179                 As a general rule, copyright owners are opposed to statutory licenses. Some, I am sure, will object that the Digital Retailer model is a statutory license, and be suspicious of it or even hostile, for that reason. Their objections ought to be soothed by their freedom to set their own royalties. But some copyright owners may say that the Digital Retailer model – tied, as it is, to online redistribution –simply will promote unlicensed (and uncompensated) CD and DVD burning, and a return to the “sneaker net” of disks and tapes that “were handed in person between members of a group or were sent by postal mail.”[34] Indeed, it may; and if it does, the issue of blank media levies will take center stage again.

180                 Today, though, I am more concerned that the Digital Retailer model will draw objections from consumers and their advocates. (Hardware and software companies are relieved of all burdens by the Digital Retailer model. And ISPs should be satisfied – indeed, pleased – with their share of the retail take.) My concern is that even if the implementing technology works perfectly, so consumers are charged only for what they choose to buy and only at prices they’ve agreed to pay, they or their advocates will view the Digital Retailer model as one that gives copyright owners too much control.

181                 Any argument that copyright owners would have too much control is one that would overstate the extent to which copyright owners have exclusive rights to their works. Copyright law gives copyright owners very thin protection. It does not protect ideas or concepts[35] or theories or the facts on which they are based.[36] This means that although “Mickey Mouse” belongs to Disney, “Mighty Mouse” does not. “Mighty Mouse” belongs to Viacom.[37] And while copyright law does not permit others to make exact copies of “Mickey” or “Mighty Mouse,” it does permit unregulated breeding of other animated mice, by all who wish to do so. It also means that copyright gives no one the exclusive right to tell stories about archaeologists in search of artifacts hidden in snake-infested caves while simultaneously confronting dangerous human antagonists. Anyone who wants to tell that story, may.[38]

182                 In the music business, copyright law doesn’t give record companies the ability to obtain exclusive recording rights to songs. Instead, all who want to are permitted to make and sell their own recordings of popular songs – even sound-alike versions – simply by paying license fees to music publishers at rates set by law.[39] (MP3.com could have started its own record company – producing sound-alikes or original recordings – for less money than it agreed to pay in settlement of copyright infringement lawsuits filed against it by record companies.[40])

183                 There has been debate over the appropriate scope of the derivative work right (the right to make new versions of copyrighted works).[41] But the derivative work right is not at the heart (or edge) of the digital copyright controversy. Digital copying and online redistribution involves exact duplicates of copyrighted works.

184                 Given – and this is a “given” – that copyright law permits anyone to breed new animated mice, tell new stories about adventuresome archaeologists, and make new recordings of “Oops! I Did It Again” using vocalists who sound just like Britney Spears, it hardly seems too much to ask that they do so, rather than make unauthorized digital reproductions of works whose copyrights are owned by others.