Is this company censoring reports of Ukrainian Nazis on Twitter?
Is this company censoring reports of Ukrainian Nazis on Twitter?
Shadowy "fact check" site Logically says it uses AI "to sift through information." | Issue #46
I was censored by Twitter on Thursday. They prevented me from posting a factual tweet about Ukrainian Nazis. The tweet was auto-hidden, and am locked out of my account (banned from posting new tweets) until I agree to click “delete” on the hidden tweet.
You can read the banned tweet and more details of the censorship in my post from Thursday. But now I suspect I may have found the source of the ban.
Many thanks to everyone who’s been sharing my post to Twitter over the weekend. Please keep up the pressure: keep posting the link to Twitter, keep mentioning @Twitter, @TwitterSafety, @TwitterComms and ask them to reverse the ban.
The speed with which Twitter locked my account and hid the tweet (it happened instantly) suggests that the ban was done automatically — using an algorithm. I had assumed it was one of Twitter’s own automated processes, but it’s starting to look like the decision may have been farmed out to a contractor.
Researching that article, I looked deeper into the origins of the four photos I had attached to the banned tweet.
I did reverse image searches on the photo above, which shows two Ukrainian fighters in combat gear, with one wearing a Black Sun Nazi symbol.
That in turn led me to a page on the website of a company called “Logically.”
The page is a strange little “fact check” of the photo of the two women.
The page states that many of the claims about the photo posted online are “false” and that it is not even a recent photo.
But the page gives no sources (apart from two links to general background information about the Black Sun).
It does claim that “a reverse image search by Logically revealed that many of the photos are actually from several years ago,” and that this photo “is actually from at least four years ago”.
But I could not replicate these findings.
The page claims that the photo “can be found appearing in posts from as far back as 2017” but gives no links or other evidence for this claim, for example by explaining where this photo was allegedly posted online in 2017.
When I performed reverse image searches using Tin Eye and Google Images, neither engine found copies of the image older than a few weeks.
The oldest posts I could find were these from Reddit, and Twitter on 27 February 2022:
Tin Eye did turn up two false positives: entirely unrelated 2019 and 2017 articles about the Unites States on a Bulgarian website which — judging from the Tin Eye listing — appear to have linked to and displayed a thumbnail of an article about the two Ukrainian troops sometime shortly before 1 March (it is common for content on websites to automatically load previews of newer articles.)
In any case, Logically’s claim that the photo is “false” is highly misleading. Official Ukrainian sources later confirmed the photo is genuine.
On International Women’s Day, the official blue tick Twitter accounts of the Ukrainian Parliament and the foreign ministry both posted the photo along with the text: “women make up more than 15% of the regular Ukrainian army” (suggesting the woman in the photo wearing the Black Sun Nazi symbol is indeed part of the regular Ukrainian army, and not a volunteer as suggested by the fact check).
The point is moot anyway: the Ukrainian government officially coordinates volunteers, and the Nazi Azov militia was integrated into the armed forces way back in 2014.
Logically did not respond to emails requesting comment and asking whether they are Twitter contractors. They also did not respond to a fact check correction which I submitted using the form at the bottom of the page.
The company describes itself as combining “advanced AI, expert OSINT investigators and one of the world’s largest dedicated fact-checking teams to fight damaging misinformation.”
“OSINT” is trendy jargon for “open source intelligence” — i.e. finding things on the internet. It’s a term favoured by Bellingcat and other shady intelligence contractors close to US and UK spy agencies.
In an interview with The Times of London, Logically founder Lyric Jain explains that the company is based in West Yorkshire and in India and uses an “assembly line” for fact checking.
His company combats governments it says are running “black-hatted marketing campaigns,” including Russia and China. According to the journalist who interviewed Jain, the CEO has unnamed “foreign governments asking for his help.”
According to a second interview with Jain, “Logically’s use of AI sets it apart from other fact-checking businesses, which typically rely on humans to sift through information to determine a story’s validity — like Snopes, a US-based fact-checking platform.”
By way of contrast, “Logically’s reliance on AI speeds up the process.” According to Jain: “If claims have been checked before, a result can be given almost instantly.”
This appears to have been what happened in the case of my tweet. But in light of the poor standard of this particular “fact check” by Logically, they should not be given power to prevent free speech online.
Hmm.. A quick check on the status of the company 'logically' seems to indicated that it's a private company in it For Profit, like a gun for hire... Hmm...