Misinformed reports on misinformation

The headline is absolutely wrong, as is the opening paragraph.

BuzzFeed has obtained a statement from Facebook in which the tech giant admits, for the first time, that some Russia-linked accounts may have used its platform to try to interfere in the UK’s European Union referendum vote in June 2016.

However, if we follow the BuzzFeed link, things get interesting.

Responding to two questions from BuzzFeed News on Monday about whether there were any Kremlin-linked ads on Facebook around the time of the 2016 Brexit vote, a spokesperson said the tech company had not "observed ... significant co-ordination".

“To date, we have not observed that the known, coordinated clusters in Russia engaged in significant coordination of ad buys or political misinformation targeting the Brexit vote," said the spokesperson. When pushed about whether the statement contradicted a top Facebook executive's earlier words that were was no evidence that Russia interfered in Brexit, a spokesperson said the official statement about "significant co-ordination" was the one the company was standing by.

OK, not so bad. "We haven't observed it" = "no evidence that it's happening", fair enough (which is very distinct from "evidence that it is not happening"). But then we get :

The company refused to clarify whether the statement meant it was aware of at least some coordinated action during last year's referendum, despite several requests.

Benefit of the doubt : Facebook put up a minor lackey as a spokesperson who wasn't authorised to say anything at all beyond the official statement. But it will be very interesting to see how this one develops. Especially, as the original BuzzFeed article goes on to correctly point out :

While its claim not to have found “significant coordination” of Russian activity ahead of the Brexit vote might sound like ‘case closed’ on the EU referendum front, the company has consistently sought to play down the impact of Facebook-distributed Russian misinformation — with CEO Mark Zuckerberg initially describing it as a “pretty crazy idea” that fake news could have influenced voters in the US election. Nearly half a year later, after conducting an internal investigation, Facebook conceded there had been a Russian disinformation campaign during the US election — but claimed the reach of the operation was “statistically very small” in comparison with overall political activity and engagement.

Finally, at the end of last month, about a year after its CEO’s denial of the potency of political disinformation on his mega platform, Facebook admitted Russian-backed content could have reached as many as 126 million people in the US. It now estimates the number of pieces of divisive content at 80,000, after being asked by congressional investigators to report not just direct Russian-bought ads but organic posts, images, events and more, which can also of course become viral vehicles of disinformation on Facebook’s algorithmically driven platform.

So there’s a reason to be cautious about accepting at face value the company’s claim now that Russian Brexit meddling existed on its platform but was not significant.

NO ! Dammit TechCrunch, they did not make that claim. A statement that "no significant co-ordination was observed" is neither mutually exclusive with nor supportive of the claim that "some significant co-ordination was observed". That's not how language works. Facebook's refusal to deny that any co-ordinated action was observed is very interesting, but that's all it is. For now.

That said, I wouldn't be the slightest bit surprised if this ties up with Cambridge Analytics, Mercer, Nigel Farage, etc. In fact I'd bet that it probably does. But this is based on gut instinct, not rational, objective evidence.


https://techcrunch.com/2017/11/14/facebook-says-russia-did-try-to-meddle-in-brexit-vote/