From the Inside: A Twitter clean-out casts light on the world of sockpuppets, bots, and Putin-Zuma thugocracy

Friday, 13 July 2018, started out like any other day, until I opened my Tweetdeck.

Scrolling down my Twitter timeline, I was immediately struck by the altered tone. Gone were the vicious attacks, the deliberate distortions, the vitriolic insults, to which I have become inured over many months.

Instead, there was (mostly) civil engagement, even from people challenging or disagreeing with me. For the first time in years, I experienced Twitter for what it always had the potential to be – a platform for open debate.

It was a pleasant surprise, but also slightly puzzling. What had changed? And why?

Then I recalled that, four days earlier, Business Day had reported that Twitter was in the process of cleaning out “fake and dubious accounts” in an “aggressive bid to stem the flow of false information”.

They were doing so for two reasons: a) to stop the practice of people buying fake followers in order to appear more influential than they actually are, and b) to stop the use of “sockpuppets” and bots to manipulate public opinion.

For readers unfamiliar with the vocabulary of social media, a sockpuppet is a fake online identity created for the purpose of deception, often to influence public opinion in order to drive a hidden agenda. A sockpuppet is different from a pseudonym (which people often use to mask their identity, while still writing as themselves). A sockpuppet, in contrast, establishes a separate, fictitious identity for ulterior motives.

A bot is a fake account, activated by a computerised algorithm (rather than a person), often with the intention of multiplying the reach of a sockpuppet’s posts. The two often work together.

The devastating impact of the combined use of sockpuppets and bots in “black ops” campaigns entered general public consciousness earlier in 2018 as a result of the Mueller indictments, in the United States. Former FBI Director Robert Mueller, who headed the inquiry, detailed how un-attributable social media campaigns were deliberately harnessed by Russian agents to polarise public opinion in the run-up to the 2016 election in the United States.

The indictments expose how a tightly knit cabal of just 13 people mobilised a network of fake accounts (sockpuppets and bots) to reframe the public debate in a country of 323-million people. Behind the façade of Twitter handles and Facebook profiles, this small group drove a digital divide-and-rule strategy, creating fictitious personalities, whose posts were carefully designed to cleave open the fault-lines of American society.

Instead of spreading the benefits of open dialogue (as many analysts automatically assumed social media would do) their global reach has shown just how vulnerable democracies are to hostile manipulation, masterminded by autocrats who understand the power of algorithms. When these strategies harness Big Data and “artificial intelligence” the impact could be devastating for the world’s democracies.

This week the Mueller revelations continued, as we learnt that a 12-person team of Russian operatives was indicted for hacking into email accounts of the Clinton campaign through a technique called “spearphishing” – tricking one key person into giving away password details and using this computer as a gateway to others.

The extent of the abuse that emanated from Russia’s benignly-named “Internet Research Agency” has created a “free speech” dilemma for all social media platforms. Should fake accounts, hiding behind false identities, be free to influence public opinion to advance political agendas?

Twitter’s answer, fortunately, is: No. The problem had become so pervasive, it was threatening Twitter’s business model. And this is primarily what motivated the clean-out.

As the New York Times put it:

“The reform comes at a critical moment… Twitter has been sharply criticised for allowing abuse and hate speech to flourish on its platform. And along with other social networks, Twitter was a critical tool for Russian influence during the 2016 election, when tens of thousands of accounts were used to spread propaganda and disinformation. Those troubles dampened Twitter’s prospects for acquisition by a bigger firm, and the company, which went public in 2013, did not turn a profit until the final quarter of last year.”

In other words, if Twitter wants to remain in business, it will have to curb this abuse.

As I considered these factors, I asked myself: Could the Twitter clean-out have been responsible for the dramatic reduction of abuse on my own timeline?

Sure enough, when I checked my profile, it turned out I had lost almost 25,000 followers overnight. According to news reports, this was one of the largest drops on any South African account – roughly the same as Julius Malema’s, whose timeline had been populated by a similar number of sockpuppets and bots as mine.

The number of followers we each lost represents only a small percentage of the total. But, speaking for myself, it must have constituted a significant proportion of those responsible for the toxicity on my timeline. The drop of 25,000 followers had an enormously positive impact on the tone and quality of the interactions.

(Interestingly, by Sunday morning, my follower count had risen again by about 600 and the venom was beginning to return. It is a trend I will be monitoring closely.)

But the implications of this issue extend far beyond the personal.

It is in every South African’s interest to demand more information from Twitter. What are the “handles” (names) of the deleted fake accounts which infected South African timelines? Who launched them? From which geographic locations and when? Answers to these questions will bring us closer to answering the crucial question: Why?

Some people may have “bought” fictitious followers to fake the extent of their reach, but that is a side issue. In most cases, fake followers are unwelcome intruders on anyone’s timeline. They find their own way there, for another reason.

For a while now, I have been working on a hypothesis, which was strengthened this week by a serendipitous meeting with two international experts.

The two researchers, from a reputable international foundation, have spent their professional lives studying the methods used by Russia to influence international developments. The two men are well acquainted with the information emerging from the Mueller indictments and have now turned their interest to South Africa, partly because of the close ties that developed over the past five years between Vladimir Putin and Jacob Zuma.

Both Putin and Zuma have a lot in common, and according to all accounts, established an instant rapport at their first meeting, shortly after Zuma was elected president for his first term (Putin for his third).

Soon after their initial meeting, their interactions jumped the tracks of “official channels” and Zuma paid several personal visits to Putin, free of the constraints of state protocol.

They understood each other. Both men had reached the pinnacle of power after spending most of their professional lives as high-ranking intelligence officers.

Putin graduated from the KGB to head the Russian Federal Security Service. Zuma directed the ANC’s intelligence operation during its years of exile. Both mastered the art of patronage to reward the loyalists who kept them in power; both developed close links with criminal syndicates to enrich themselves and their network. Bratva (the Russian word for Mafia) even rhymes with Gupta! The criminal networks in both countries sought to capture the state through exercising undue influence over their political patrons.

On a recent visit to South Africa, another scholar, Dr Nancy Ries, who teaches Russian Studies at Colgate University in the United States, traced an identical pattern, and called the resulting system a “thugocracy”.

Apart from protecting its allies, a thugocracy must destroy its enemies, and their primary tool is the manipulation of communication networks to divide and weaken their opponents.

As the many layers of Russia’s strategy to divide and polarise American society are peeled away by the Mueller indictments, I am increasingly convinced that a proper excavation into the murky depths of the Zupta saga would reveal a similar modus operandi here.

Just this weekend, Public Enterprises Minister Pravin Gordhan warned that “Bell Pottinger 2” was well under way to “push back investigations into State Capture”.

Bell Pottinger part 2, Gordhan said, “is an underground media strategy to attack those who stood up against wrong things”.

So what has this got to do with the Twitter clean-out?

It provides a significant additional dot, which, when connected with many other dots, reveals a bigger picture that has been emerging over many months.

Joining dots is a crucial starting point for investigative journalism. When there are enough dots, it is possible to formulate a hypothesis, based on a series of observations, which must then be tested rigorously, until it is either confirmed (in whole, or in part) or rejected.

I have used this method before, as I did here in 2014, long before details of the “capture” of the South African Revenue Service were revealed. The media at the time where dominated by leaks from the Zupta propaganda machine for the purpose of planting misleading media coverage. By joining the dots back then, I wrote up the alternative account. Much of my hypothesis, back then, has been confirmed.

So I set out another hypothesis again today, based on what we learnt over the past year, including my conversations this week, culminating in the Twitter bot-cull.

1. The connection between Zuma and Putin was much deeper and broader than we understand, and involved their broader networks. These have not ended. Indeed, they continue to operate.


2. The nuclear deal was just a symptom of their close co-operation and shared modus operandi in partnership with criminal syndicates to milk the state.


3. If we want to understand their methods and strategies, we should be following the Mueller indictments very closely.


4. The methods of communication – including the fraudulent abuse of social media – have been pivotal to both strategies (and continue to be in the South African context).


5. The modus operandi included the systematic discrediting of people (public officials, journalists and politicians) identified as a risk to the network.


6. In South Africa this involved the capture of key media groups, and abusing state resources.


7. Social media was the most important communication platform underpinning this strategy, with a plan to distort information and discredit individuals, by manufacturing outrage.


8. Backed by the reach and thrust of social media, these campaigns were given additional volume through radio, print and television.


9. The plan aimed to rip open the cleavages in South African society, and polarise communities, primarily on the basis of race.


10. This campaign continues in order to protect and entrench the Thugocracy.