Sunday, October 14, 2007

Barron tribute, panel 3: me

My presentation:

People these days often use the First Amendment to say something about copyright; I’m going to use copyright to say something about the First Amendment. Even in a world of user-generated content, intermediaries play a key role – American Idol has an ideology of amateurism, but is actually scripted, edited and influenced by the network and the advertisers; singers sing preexisting songs rather than using their own compositions. How does this relate to access to the media? Any access policy has to address the second thing people do when they get access to a new communication channel, which is to copy (the first being to create porn, and indeed the early internet cases often involved copying porn).

On the internet, there was a relatively early response, the DMCA, limiting intermediaries’ liability for monetary relief when users post infringing content and establishing a notice-and-takedown regime. This was important because the standard secondary liability doctrines did not have safe harbors, though in many ways the DMCA replicates traditional secondary liability. There are strong critiques of the way the DMCA enables copyright owners to suppress criticism and other content, but I will not dwell on them.

My interest in the DMCA is in what it’s not. No one, as far as I can tell, suggested that it was constitutionally required. When we talk about government neutrality in the marketplace of ideas, the legal treatment of intermediary liability reveals that to be an incoherent concept; the legislature has substantial freedom to shape speech by regulating intermediaries. Compare the DMCA (safe harbors, encouraging compliance with copyright owners’ complaints) with the CDA (no liability to people hurt by users’ content no matter what, plus immunity from claims by users whose content is suppressed by voluntary content policies) with current actions involving internet users convicted of sex crimes (MySpace and other social websites under investigation and threatened with new regulation because of concern they enable child predation). The legislature and executive are apparently free to act, as long as they don’t ban individual speech.

Here’s an alternative constitutional narrative, beginning, like much modern First Amendment law, with NYT v. Sullivan. Sullivan was a case about intermediaries and specifically about protecting the business model of the NYT, which was to publish others’ speech. Why wasn’t a negligence standard enough to protect the Times? The Times was too distant from the facts, and thus should have known it needed more verification. As a large organization, indeed, it had more resources to verify truth, including stories within its own files, a point the Alabama Supreme Court emphasized in its decision upholding the verdict. The NYT also had less incentive to host others’ words than its own, and as a profit-seeking entity was more likely to be deterred from speaking than an individual printer distributing speech he fervently believed to be true, who might overestimate his chances of avoiding a libel suit or judgment. A liability standard sufficient to prevent chill of a yeoman speaker would have too great a deterrent effect on an intermediary like a publisher or bookseller.

These considerations are all apparent in the Court’s opinion (a sharp contrast to the concurrences, which perhaps in response emphasize that they’d treat individuals and the press the same in barring all defamation claims by public officials), but they’ve been lost in the general understanding of Sullivan as an individual right. Concepts like clear and convincing evidence and actual malice are much more useful for intermediaries like the Times – which not incidentally have deep pockets are are more attractive to sue, as the Court pointed out – than for individual speakers, who are less chillable.

If you accept this version of Sullivan, some constitutional protection for online intermediaries might seem to follow. If Sullivan is not enough for AOL because it can’t even give its content the screening the Times can (a proposition strongly urged by ISPs and adopted by Congress in the CDA), and if AOL is a publisher entitled to First Amendment solicitude, then we need a super-Sullivan. (As Jack Balkin pointed out to me later, Google’s indexing policy defines reckless indifference for the purposes of the actual malice standard – Google just does not care whether the information in a webpage is true or false.)

But the DMCA and the CDA ended any moves in that theoretical direction. Imagine Napster or Grokster making the Sullivan argument that their business models couldn’t survive if they were liable for the content whose transmission they enabled.

From the intermediary side, the DMCA tells us that the legislature has a lot of leeway; Congress could constitutionally repeal the DMCA, despite the allegedly crushing effects that would have on internet intermediaries, or constitutionally legislate greater burdens on them to screen speech.

What we’re left with is intermediary power without much responsibility. Content screening is okay – and being embraced by big aggregators like YouTube – even though it’s not mandated. In the CDA, the liability/power gap is even greater, enabling huge amounts of private power over individual speech without responsibility – Section 230’s title is actually “protection for private blocking and screening of offensive material.” The recent case is an example of the unchecked power this creates.

One reason to be concerned with intermediary power in copyright is that copied speech is often important to get others’ attention, as Netanel discussed, regardless of whether it comments on the original. But the broader points are theoretical – looking at intermediary liability makes clear the enormous power of the state to shape speech, favoring copyright owners and disfavoring victims of defamation, in our current version – and perhaps policy-oriented – we should think about a comprehensive intermediary liability standard, and about implementing nondiscrimination norms as a quid pro quo for immunity from liability based on user-provided speech.

No comments: