Filtering Free Expression

On August 15, the National Telecommunications and Information Administration, a division of the U.S. Department of Commerce, released a report that reads like a sales pitch for Internet filters.

Although the report conscientiously outlines the massive, well-known over- and under-blocking problems with filters, including the extensive findings of the trial court in the recently concluded "CIPA" case, it concludes that the defects are inconsequential, and that the real problem lies with teachers, librarians, and administrators who simply need more training in how to use filtering technology.

The report was a response to a mandate from Congress in the 2000 CIPA law (the "Children's Internet Protection Act") to "evaluate the effectiveness of Internet technology protection measures" -- that is, filters. CIPA requires such filters in all computers -- for both minors and adults -- in schools and libraries that receive e-rate discounts or other forms of federal aid for Internet connections.

The NTIA waited until after the Supreme Court ruled on the American Library Association's constitutional challenge to CIPA before releasing its report. Despite the findings of the trial court in the CIPA case that filters, even at their narrowest settings, block tens of thousands of valuable Web pages on religion, public health, and countless other topics, the Supreme Court upheld the controversial law in June 2003.

The NTIA report relates that the agency received comments from educators outlining the difficulties with filters. It quotes one educational study: "Most of our students feel that filtering software blocks important information, and many feel discouraged from using the Internet by the difficulties they face in accessing educational material.” Yet the report accepts unquestioningly the submissions of filter manufacturers that tools are available (including disabling the filters) to solve the problem.

For example, the report says that "several vendors' comments discussed their products' ability to allow users to type in a Web address to learn more about a particular site's blocking category" -- without explaining how this information would restore erroneously blocked sites to a student's or teacher's computer. It accepts at face value the claim of one company that it reviewed more than 60,000 requests to "add, delete, or change a blocked Web site" during a 7-1/2 month period; then fails to discuss how this asserted responsiveness to public pressures or complaints would lead to an objective or rational system.

The disconnect between facts and conclusions is most evident in the section of the report titled "Fostering the Development of Measures That Meet the Needs of Educational Institutions." Here, the NTIA outlines five features that would improve filters, including "image recognition technology" and "technology that scans a Web site's content, rather than relying on key words.” It fails to note that attempts to develop "image recognition" software that can distinguish human nudes (or human nudes engaging in sexual activity) from other images have been notoriously unsuccessful; or that, as the CIPA trial court found, there is no way that technology, unaided by human judgment, can really evaluate Web content. Nor does the NTIA mention that given the size of the Internet, with millions of sites changing daily, there is no way that humans can review and make judgments about every page on the Web.

Instead, the report naively accepts the claims of filter manufacturers that they already offer many of the features desired by educators -- including the incredible assertion by one vendor that it relies on "100 percent human review of all Web site content," and the obfuscatory language of another asserting that its "comprehensive content monitoring" tracks "all inappropriate content flowing through the network."