On 30th August 2017 The Times ran an article on its front page saying “evidence suggests” the unverified Twitter account @DavidJo52951945 (now @DavidJoBrexit) is “a suspected Russian troll posing as a UKIP supporter”. It was largely based on a Twitter thread by an anonymous data scientist, @Conspirator0
Here's an interesting observation – David is posting 8 AM – 8 PM every day, Moscow time. Almost like it's his job or something. pic.twitter.com/qqRECn14qj
— Conspirador Norteño (@conspirator0) August 23, 2017
I highly recommend you read the thread to see the full analysis, but the main elements of the argument focussed on:
- timing of the account’s tweets claimed to correlate to Moscow office hours;
- the flow of subject matter and messaging tweeted over time;
- account username ending in an 8 digit number being indicative of bot or troll accounts (probably because it’s a format of standard suggestions when you create a new username); and
- the account being a key node in a network of 63,000 accounts with similar 8 digit usernames.
Not everyone was convinced. Some examples being:
Alex Hern, Technology features writer for the Guardian expressed his scepticism in this thread.
I’ve also been in touch with one person – a UKIP councillor – who says he’s met the man who posted the ballot picture.
— Alex Hern (@alexhern) August 31, 2017
the account @hughster provided some alternative evidence in this thread
However, you might be interested to see how David Jones started out on Twitter: spamming an anti-immigrant petition. https://t.co/3q3XIjjCuA
— Hughster (@hughster) August 24, 2017
1. Since The Times did its Twitter-scrape story marking @DavidJo52951945 as a 'Kremlin agent' there has been an unfortunate consequence.
— Mike Hind (@MikeH_PR) September 1, 2017
There is evidence showing @DavidJoBrexit is potentially an influential account:
- he ranked 14th in our recent 2017 General Election analysis based on influential tweets analysed by our Tweetsdistilled application and featured on our @PoliticsUKTD account;
- ranks 5th in Lissted’s UKIP community (after Nigel Farage, UKIP’s party account, Robert Kimbell aka @RedHotSquirrel and Patrick O’Flynn)
- his tweets have been used to illustrate articles by MSM e.g. Telegraph, Huffington Post, BBC and (perhaps surprisingly) even The Canary; and (as of 4th September 2017)
- he’s followed by high profile figures including Eamonn Holmes, Leave EU, Kate Hoey MP and prospective UKIP leadership candidate, Anne Marie Waters.
Meanwhile his bio and profile tell you virtually nothing about who he is. None of this tells you if he’s a Russian agent or not. But it does provide evidence he’s been using Twitter as a platform to build influence and we know virtually nothing about who he is.
According to The Next Web, there are around 270,000 verified accounts on Twitter. This is where Twitter has carried out various checks to establish the identity of the person tweeting. Followers then have confidence the account is genuine and it also provides some protection from passing off by others for the individual concerned.
Twitter initially only verified accounts relating to celebrities, politicians, top brands, key public sector bodies and some high profile journalists. It then began to expand who it verified and last year even opened up the process so that anyone can request to be verified, although that doesn’t mean they will.
Other Unverified Accounts in PoliticsUKTD GE2017 Analysis
DavidJo52951945 was actually one of 26 unverified accounts to rank highly in our analysis of tweets in relation to the 2017 UK General Election.
The other were (followers in brackets):
- NHSMillion (227,616)
- britainelects (156,560)
- V_of_Europe (152,230)
- JeremyCorbyn4PM (140,459)
- davidallengreen (113,178)
- ToryFibs (88,751)
- ShippersUnbound (73,701)
- AngryScotland (previously AngrySalmond) (73,061)
- JamesMelville (64,877)
- hrtbps (63,401)
- RichardJMurphy (45,157)
- Rachael_Swindon (44,554)
- MSmithsonPB (43,455)
- PickardJE (40,016)
- AaronBastani (37,396)
- mikegalsworthy (33,059)
- imajsaclaimant (32,482)
- RedHotSquirrel (31,677)
- thepileus (31,145)
- EL4JC (17,673)
- DavidPrescott (12,002)
- GAPonsonby (11,546)
- ReclaimTheNews (11,283)
- BrianElects (8,007)
- DaveWardGS (6,806)
These accounts are from across the political spectrum. They include some where it’s obvious who they relate to e.g. Jim Pickard and David Allen Green. But they also include ones where, like DavidJo52951945, the tweeter is either completely anonymous or there’s very little information about them readily available.
I asked a couple of the most high profile accounts – Jim Pickard and David Allen Green – if they’d been approached by Twitter regarding verification. Both confirmed they didn’t believe they had.
Don't think so as far as I know
— Jim Pickard (@PickardJE) September 8, 2017
Can confirm never been approached. My DMs open if you want to discuss.
— Law and policy (@davidallengreen) September 8, 2017
High reach unverified accounts
So we’ve seen that there are influential accounts that aren’t verified. Even ones with a high reach of 100,000+ followers.
We searched in Lissted’s database of potentially influential Twitter accounts to see if we could find more. We used criteria including: 100,000 or more followers: tweeted since 1st June 2017: tweets public and have posted at least 10 times; and time zone relating to North America, Europe, Australia or no time zone is stated.
As of 4th September 2017 this produced a list of 14,710 accounts. In aggregate these accounts have 6.1 billion followers.
Just because an account has a high follower number doesn’t mean it’s influential. And vice versa. There are many more accounts with smaller reach that we would consider influential as we saw illustrated with the GE2017 analysis.
However for this exercise we’ve taken this highly simplistic approach because it provides a degree of consistency with the reach of DavidJoBrexit*.
What sort of accounts are on the list?
There’s a diverse range of accounts on the list. Celebrities e.g. @AlanCarr and @Daraobriain. Meme e.g. @GirlPosts and @SoDamnTrue. “Pics” e.g. @Earth_Pics and @HistoricalPics. Cats (of course) e.g. @EmrgencyKittens and a whole host of others including the four accounts in the GE2017 analysis with 100,000+ followers. Many of the accounts will just be tweeting harmless content. I’d guess a good number have a high fake follower quotient.
But are they all harmless? What if like DavidJoBrexit some of these accounts are sharing content of a more significant nature and their followers include a lot of real people who are actively engaging with the account and its content? What if some of those followers are influential in their own right?
In such situations is it important that followers know more about the identity, nature and potential motivations of whoever is behind such accounts?
Does Twitter need to take action?
Twitter’s verification activities have addressed a large number of high profile and influential accounts, but it appears from this research there are still a significant number who remain unverified.
There’s an important balance to be struck here between the comparative risks and benefits of verification and anonymity.
As mentioned above David Allen Green, lawyer and journalist (and one of the most influential UK accounts on Brexit and legal matters in general), hasn’t been approached by Twitter to be verified. He also hasn’t requested to be verified himself. He was kind enough to share his views on verification with me.
“I have not sought a blue tick for three reasons. First, it does not cause inconvenience. The lack of a tick does not affect how I use Twitter. If this changes, and it becomes inconvenient, then I may well resign myself to getting one.
Second, the methods I use to assess the credibility of an account do not include the tick. There are some outstanding accounts with no blue tick and some dreadful ones with one. I prefer to use peer recommendation, shared followers, tone of timeline, and use of supporting links.
Social media can be egalitarian, and many good accounts rise by merit alone. Some of the best accounts on Brexit, for example, are under pseudonyms – but the content is verifiable and spot-on. A caste system goes against this.
Third, blue ticks are a form of regulation, and social media is better when there is less regulation. Yes there is vile nastiness, but there are also valuable things being said every day which would not otherwise be heard. I am not suggesting there should be no regulation, but there is the risk that good things are lost when you try and ban bad things or impose a tiered system.”
These are good arguments, but are they all applicable for the vast majority of Twitter users? Most users spend very little time on the platform and I wonder how many are as diligent and expert in their assessment of an account and its content as David?
If Twitter was to require all influential accounts to be verified this would clearly present some challenges where anonymity is important and beneficial.
There’s also the question of where to draw the line on what constitutes “influential”? And if influential account owners aren’t prepared to be verified then should Twitter go as far as to consider suspending their access?
These are tricky questions. But when suggestions are made about the impact of social media on major real world events is it a greater risk to allow influential accounts to operate either anonymously or without followers knowing if the user is who they claim to be?
This is an edited version of the original post. The full post can be found here.