MisDisMal-Information Edition 8 – This Week’s Information Disorder Letter

This newsletter is published at techpolicy.substack.com

What is this?This newsletter aims to track information disorder largely from an Indian perspective. It will also look at some global campaigns and research
What this is not?A fact-check newsletter. There are organisations like Altnews, Boomlive etc who already do some great work. It may feature some of their fact-checks periodically

Welcome to Edition #8 of MisDisMal-Information

Information Operations #LadakhBorder and #WorldWar3

The New Indian Express ran a rather curious story claiming that ‘hundreds of fake Twitter accounts from Pakistan’ were spreading lies.

They have since updated the headline to read Ladakh standoff: Hundreds of fake Twitter accounts from Pakistan spread lies in Chinese garb.

It cited some research from Technisanct and quoted the CEO as saying

“As Chinese-Indian tensions started rising, we observed a huge growth in retweeting of pro-Chinese tweets. We identified multiple Pakistan operated handles that started to change their names and translate tweets into Chinese. Most of the accounts have a Pakistani flag and Chinese flag in their handles and a bio to create a feeling that Pakistan is highly backed by China,”

I am not entirely sure what this means, and I wasn’t able to find a full report on the company’s website. They do have a section for Case Studies and Press Releases, neither mentioned this.

The Technisanct team used a platform called Twint and a trends map to gather information related to the discussion that happened in the aftermath of the Ladakh issue. After observing these Twitter accounts, its followers and past tweets, it was learnt that these are operated to propagate pro-Pakistan narratives. This is a strong organised activity working in a structured format creating Twitter groups and appointing administrators who get instructions from top-level individuals, he said.

Now, let’s be clear, such events are opportune times for adversarial actors to run information operations. And this impersonation is very common. We’ve seen this tactic deployed domestically as well as by foreign actors, What concerned me about the article was a lack of clarity around certain elements:

1) Is there a detailed/full report. I really think publications should start linking to full reports or insist on more clarity before running them.

2) How were ‘fake’ accounts classified?

3) How did they conclude that it was organised activity? Is the ‘structured format’ the only signal that was used to determine this?

4) Apart from impersonation, what other false information were they spreading? There were 1-2 anecdotal examples, but nothing else.

5) Is pro-Pakistan narrative the same as anti-India narrative? Depending on the context, that may or may not be the case. My point is that it just isn’t clear.

This is an emotive subject and I am coming at this from an academic perspective. Many questions around ‘coordinated’ activity, ‘fake’ accounts are still open, so these terms shouldn’t be bandied about loosely. But as Thomas Rid points out in Active Measures, overstating the impact of such operations is dangerous as well.

Anyway, I did pull tweets related to both these hashtags at around 12 PM on 20th June. Starting 18th June, there were a total ~23K tweets on both hashtags (~17.5K on #LadakhBorder and ~5.5K on #WorldWar3). I ran the pull operation twice to double-check this was as all there was, turns out, it was. This is certainly not a high number

I also looked at a subset which came from handles that explicitly set the location as Pakistan (~1200 tweets). Some statistics:

Number of Users who tweeted:

  • Full Dataset – ~18000
  • Subset – ~ 780

Total Likes (as reported by Twitter’s API)

  • Full Dataset – ~52000
  • Subset – ~ 3200

Full Dataset

Subset

Full Dataset (Count of Tweets on the right axis)

Subset (Count of Tweets on the right axis)

If you’ve been following similar analysis in this newsletter, there isn’t much that stands out in any of the 4 charts so far.

The only thing that struck me was the fact that the accounts with the most number of tweets in both datasets matched. When I looked into them, they both appeared to be handles that share news – nothing more. Since they are not related to an individual, I can include twitonomy links for both of them.

BaaghiTV

S__u__n__Shine

I also looked at the Tag Clouds (excluding #LadakhBorder and #WorldWar3).

….

Go here for the complete edition.