Data scientist Jeff Kao says that more than a million public comments in support of the FCC’s planned rollback of net neutrality rules are likely fake. Kao also, by removing the fake pro-repeal comments, concluded that more than 99% of the unique, human-authored comments on the system favored maintaining current net neutrality rules, which prevent internet service providers from charging more for some kinds of data than others.
Kao used a technique known as natural language processing, or NLP, to scan more than 22 million comments submitted to the FCC’s website. He found that more than 17 million were duplicates or close parallels. But many of those were, he writes, “legitimate public mailing campaigns,” which provide boilerplate text for real people to submit.
Intriguingly, the comments that Kao ultimately concluded were ‘fake’ were actually quite diverse in their specific phrasing – but that variation was only superficial. As an example, Kao highlights the anti-net neutrality phrase “Individual citizens, as opposed to Washington Bureaucrats, should be able to select whichever services they desire.” The system used to generate the fake comments swapped out words in such phrases again and again – for instance, switching “people like me” for “individual citizens” and “products” for “services” – to produce 1.3 million superficially distinct variations on the same basic block of text.
Kao sums up the approach as being “like mad-libs, except for [political] astroturf.” And it would have been nearly impossible to spot without NLP, a form of artificial intelligence trained to understand language rather than just detect identical text strings.
Get Data Sheet, Fortune’s technology newsletter.
The proposed repeal of the current Obama-era protections would be a big win for ISPs including Comcast and Verizon, but has been strongly opposed by tech firms like Google and Facebook – though those giants are less likely to be hurt than smaller content providers. President Trump has long been critical of the neutrality rules, and his new FCC chair, Ajit Pai, has moved fast to rescind the rules.
Kao’s investigation adds to prior evidence of fake comments on the FCC system, some made using stolen personal information from real people. Another investigation found that many comments supporting net neutrality used fake emails and physical addresses, and were generated using forms such as the Electronic Frontier Foundation’s Dear FCC. Fortune’s Aaron Pressman has argued that undermining the public comment system would give a tactical edge to industry opponents of net neutrality, and that seems to be the thinking of the Trump FCC itself, which has refused to cooperate with an investigation into the fake comments by the New York Attorney General. That leaves the public comment process clouded in confusion, rendering real comments from concerned citizens politically meaningless.
We’re left with a he-said, she-said for the fake news era — and, as with most matters in a vacuum of fact, the debate has descended into tribalism. Pro-Trump outlets including Breitbart and the Daily Caller have taken up the banner of opposition to net neutrality, apparently more as a proxy battle than on the merits of the rules themselves. Small outlets like Breitbart, after all, are on a level playing field with the likes of CNN largely because of the principles of net neutrality.