DRM conference, session 3
Impacts of DRMs on flows of information
- David Wagner, Computer Science, UC Berkeley (moderator)
- Hal Abelson, MIT
- Edward Felten, Computer Science, Princeton University
- Joe Liu, Boston College Law School (paper on DMCA and research)
- Larry Lessig, Stanford Law School (paper on Creative Commons)
- John Erickson, Hewlett Packard
Hal: background: statute of Anne (big picture in background). Speaking to policy makers: watch out. You’re sounded by two dangerous delusional communities: making things better by making things more precise. Public good: legal code overwrites & fuzzifies that, IT standards overlays and fuzzifies, implementations (computer guys) on top. This is how the world is, and it’s ok. But now: mattress tag: DO NOT REMOVE UNDER PENALTY OF LAW. Policymakers: key legal principle is NOT fair use, it’s de minimus non curat lex: Do not succumb to illusion that the public good is best served by forcing the strict alignment of practices with policy. Scientific publishing: the basic deal as seen by the universities: scientist authors give to journal publishers, publishers own w/ all rights, allow scientists to retain some limited rights. ACM: I give to ACM, they allow me to post for my personal use on my web site. Elsevier: gives right to present author’s work at a conference. Chemists: can post title, abstract… NEJMed – woah! Why? helps maintain the integrity of pub process. (haha) Rights go to responsible parties that exert monopoly control because they own the infrastructure. What about indexing, extraction, other innovations in info use. Internet’s access, tools stillborn by limited access to quality sources? or stimulate network effects that lead to further concentration and monopolization of the scientific literature? DRM may exacerbate both. Legally sanctioned publishing monopolies that exert control by dominating the publishing infrastructure? and firmly entrenched by DRM tech and the law? (Implementations.)
John Erickson, HP: Policies: not code, not law? Policies are code, but increasingly not built in (good), useful where system must make choice. DRM is policy enforcement (one choice). Trusted systems on “trusted computing bases”, ideal for enforcing DRM. Good parts: may be declaratively expressed, studied, accessed, tailored to context; potential for policy creation. Nasty bits: limits by expressive language, choices by policy makers, implementations and systems subcomponents. Policy as private law: arbitrary controls w/ no constraints, do not factor in fuzzy attributes like locality or user intent, escapes for human intervention must be factored into policy and system design. People in or out of the loop. Constraining info flow: includes discovery, retrieval, use/reuse, dissemination, derivative works; constrained by opacity, policies, monitored access, etc. Encouraged by transparency and metadata, accessibility (policies, constraints, formats). Closed info spaces: built-in policies and control use (don’t work with certain browsers), proprietary formats, etc. How to challenge the code? (write and use policies in humanistic way?)
Joseph Liu: DMCA and regulation of scientific research paper. DMCA enacted 1998, liabilities for circumvention and distribution; controversial. How does this affect academic encryption researchers? should be largely untouched, everyone agrees. Exemption 1201(g) for good faith research was later developed, cases include Ed Felten, but claim was that fears were overblown. Impacts on research: “academic encryption researchers should be able to conduct some resarch w/out fear of liability under DMCA” and “DMCA will have non-trivial impact on conditions under which such research takes place.” Limits who can conduct research, imposes hurdles before research (must seek permission), limits free communication about results, avenues for publishing, requires notice and disclosure of results, affects content of published work (esp actual code, structure of exemption allows for changes via regulation mechanisms). Evaluating the impact: regulating (not infringing activities, devices) non-infringing activities and research. DMCA is not sufficiently careful about its impacts (overbroad).
Ed Felten: DRM has negative impact not only on tech and research, but also on debate. Device must be a black box that users can’t examine. Combination of tech and law makes the box black. Tech & public policy: important pp questions depend on understanding technology (especially true now for DRM), bans on understanding tech cripples debate. TIA program: PP, FBI + CIA want to mine commercial databases (security v privacy tradeoff), advocates claim “don’t worry” but need to understand tech first. Porn blocking and filtering: products claim to block only offending tech, don’t worry about overblocking. Need to see filter list to know. Electronic voting: push toward computerized voting, convenience and speed vs risk of fraud.What we don’t need to do is make process harder.
Larry Lessig: Binary blindness. Have been talking about DRM forever, need to reframe the problem. 3 kinds of people: not controlling their stuff at all, don’t touch my stuff unless you ask me first, the rest: ok to use my stuff for certain uses. Original net development ratified position of “nones” but created complaints from “all”–would shrink “some” and “all” space. Political response (“all”) to force shift in architecture away from default to support all. Now “some” and “none” lose space/power, must adopt burdonsome tech, marginalizes “none” space. Solving for the extremes: recognize the middle. Strategies, 1: courts and congress, restrict DRM to allow fair use, assure tech doesn’t displace rights; not going to work in time. 2: fight DRM, creative commons, distinguish between DRM and DRE: two different issues. Find a way to express one’s preferences, can be enforced by machines.
Larry Lessig: Binary blindness. Have been talking about DRM forever, need to reframe the problem. 3 kinds of people: not controlling their stuff at all, don’t touch my stuff unless you ask me first, the rest: ok to use my stuff for certain uses. Original net development ratified position of “nones” but created complaints from “all”–would shrink “some” and “all” space. Political response (“all”) to force shift in architecture away from default to support all. Now “some” and “none” lose space/power, must adopt burdonsome tech, marginalizes “none” space. Solving for the extremes: recognize the middle. Strategies, 1: courts and congress, restrict DRM to allow fair use, assure tech doesn’t displace rights; not going to work in time. 2: fight DRM, creative commons, distinguish between DRM and DRE: two different issues. Find a way to express one’s preferences, can be enforced by machines. Build a layer of reasonable copyright law, by expressing reasonable middle; flooding to default “some” space. Restore reasonableness thru voluntary action. Why? incentives to certain artists (increase exposure!), people can participate in debate. Identify, “I believe in (somewhat) free.” Urgency, extremes don’t control or we lose. Courts aren’t ready, congress answers to wrong dialog ($). Best interpretations of “ideals and principles” on Larry’s side, other side has “all the money in the world,” when was last time ideals won? Should be when’s the NEXT time. Reasonableness, building balance through use.
Questions: UC Berkeley student: why did Microsoft choose large content industry over consumers? panel: it’s more complicated than that… who’s controlling policies? From floor: chastity belts on tech. From floor: creative commons is in tension w/ defaults in law, publishing, existing power bases. Larry: Hal’s on CC board, so if there’s tension… Jack Valenti’s “terrorists” stealing content–one position. Millions of people define a reasonable view, can increasingly define extremists. Ed F: no conflict: CC isn’t expressing what I can do or not do. Hal: no conflict, Farber’s use of “consumer,” triggers new debate, it’s not about rights of creators. From floor: controlling data in info spaces? (no answer) Pam S: possible to develop circumvention for fair use, creative anti-circumvention laws, w/out opening pandora’s box? Jim: can try and approximate, reach reasonable compromise, how far to carry conversation or compromises. Maybe. Ed: this is one of the most important tech questions, right now: not effectively. Hal: increased access vs (attesting or ??) Joe: legal side, thought experiment: what would fair use look like in legal, regulatory environment? might resemble tax code in complexity. Public interest, people’s interaction varies. Mark Lemley: Shrinkwrap licenses, open software movement in contract sense, must look at both what big and little guys will do. CC: it’s easy to write a set of legal rules, infinite palette(?) and regulation by gov is good or bad thing, CC can facilitate DRM-not-limited-by… ? Larry: Distinction between DRE and DRM, not static, difference is dynamic effect on debate. Overhead of technology before sharing content; DRE (freer content), controlling can be inevitable. Mark L: DRE at odds w/ strong privacy protection? Larry: no, DRE doesn’t have same focus, arguments, characteristics of argument. Ed F: distinction w/ DRE: expression about permission vs enforcement mechanisms (raises privacy issues). Adon Katz?: courts, in reframing, was only way to frame or reframe question? Larry: USSC has average view of issues similar to regular people. Ordinary people don’t get it until you explain it to them. Movement is salient among ordinary people. Explain clearly and repeatedly, lot of work before we can win.