Back in 2008, I wrote in Cybersieves that “online censorship is sharply on the rise worldwide—in democratic states as well as in authoritarian ones.” Sometimes it is depressing to be right. CNET reports that major broadband ISPs in the United States are on the cusp of agreeing to a plan that would limit the on-line access of users accused of copyright infringement. The plan, negotiated between the ISPs and content groups such as the RIAA with the express approval of the Obama administration, would begin with warnings, but would progress to more severe measures such as bandwidth throttling and restrictions on what sites a user could reach. It’s not clear what would happen if a user who was under the more stringent restrictions switched ISPs, though the built-in market limitation on broadband (most consumers have, at best, two choices – cable and the local telco) limits a “market out” for those accused of infringement.The shortcomings of this approach are obvious. First, determinations of infringing activity are made by private actors, and it’s not clear what options exist for users to contest such allegations – or determinations by ISPs that they have in fact engaged in infringement. Second, it’s not obvious even to experienced lawyers and judges what constitutes infringement, or fair use. I may download a copyrighted work for uses fair or foul. Third, there isn’t much if any transparency about the standards that ISPs will use to make these decisions. Finally, how will ISPs deal with circumvention techniques, such as TOR or proxies? The RIAA has suggested recently that proxies deserve attention in this regard, which is worrisome. When combined with the PROTECT IP Act, this arrangement shows that the odds of Internet censorship in the United States are sharply up. This new private plan is more worrisome than PROTECT IP even because it is more immune to court challenge: our First Amendment rights apply only against state actors. I have a new paper in the works that offers an approach to such “soft censorship” – it looks like it’ll be even more topical than I expected. I should also point out that IP is a strange area politically: there is relatively little partisan difference in IP policy. Given the Obama administration’s lack of transparency with this initiative and with ACTA, with its policy of domain name seizures by ICE (why is an agency responsible for immigration and customs making Internet policy?), and with its support for PROTECT IP, it’s clear that the hopes that techies had for Obama were illusory. Down Under, it’s a similar story. Even though the Labor government has temporarily shelved its mandatory filtering program due to a lack of political support, several of the country’s major ISPs are set to begin voluntary blocking of URLs that the Australian Communications and Media Authority have determined to host child abuse images. Many of the policy concerns that Australian entities raised (and that I detailed in Filtering in Oz) still apply, although it’s heartening that the scheme includes annual review of the block list, standardized notification to users via a block page, and an appeals process. Overall, both of these private / voluntary schemes worry me. Restrictions on speech on the Net are even more potent when carried out by private actors – who, after all, own and operate the vast majority of Internet infrastructure and services. Moreover, these actors operate via market forces – they are accountable only to customer demand and to their shareholders. Where, however, the number of actors is relatively small, and they engage in concerted behavior, market checks vanish. And the standard protections built into public law systems – debate, transparency, periodic accountability, judicial review – simply do not exist in robust form. The failure of public Internet censorship in the U.S. and Australia was hardly the last word – it simply generated evolution in filtering, in an even more pernicious direction.