At the Digital Witness Lab, we build platforms and tools in service of public-interest research. This means a lot of our work—coding, building, and labeling—happens behind the scenes. Our WhatsApp Watch project has led us to create one of the largest sets of WhatsApp message data. To understand what we’re learning from the data, we turn to our team members working on it. Keep reading to learn from Kumar Sambhav Shrivastava, our India-based Research Lead.
The conversation has been edited for clarity.
What do you do at the Digital Witness Lab? I have been leading the lab’s effort to collect data from public WhatsApp groups in India to look at a range of social and political impacts. The team I lead collects this data, labels it to be able to make sense of what's really happening, and then looks into trends and themes to see what emerges.
Why WhatsApp? There is an immense need to bring more transparency to how WhatsApp is used in spreading political propaganda. WhatsApp doesn’t moderate the content that people share in groups because it considers them private communications. But in reality, WhatsApp’s group function has evolved into a mass communications medium. Years ago, there was a limit of 100 members per WhatsApp group. That limit has increased over time to 1,024 members per group. If you're allowing over 1,000 members in a single WhatsApp group, how can that be considered a personal, one-to-one communication channel? We’ve now observed hate, misinformation, and politically polarized content reach many people through WhatsApp groups without any sort of accountability or moderation.
Who is distributing that sort of political content on WhatsApp? We’ve found very organized, systematic influence campaigns from political parties and their collaborators in WhatsApp groups. The groups are often created by party workers, and volunteer admins help maintain the groups as part of a systematic hierarchy. The hierarchies start from the central level in a political party and break down to the state level, district level, village level, and even down to the “booth” level, which is the smallest unit of people who are designated to cast their vote at the same polling booth location.
Do political parties have any other WhatsApp strategies? Ultimately, they want to reach as many people on as many mediums as possible. We’ve seen three main approaches: Official political WhatsApp groups, seemingly organic networks that are actually run by political operatives behind-the-scenes, and truly organic community groups that political workers covertly join.
For example, we’ve been researching how WhatsApp was used by the two prominent political parties in Andhra Pradesh ahead of recent elections. We found hundreds of thousands of WhatsApp groups created by the political parties themselves. At the same time, we saw party volunteers and workers join WhatsApp groups that already existed in their communities. And unlike on more public platforms like Facebook or Instagram, political entities don’t have to identify themselves as such. So, for example, party workers and volunteers tried to access university groups, caste groups, or residential groups and then take control of what political information is shared in those pre-existing groups.
Who is joining the party-run WhatsApp groups? Over the years, many parties— like the Bharatiya Janata Party (BJP)—have developed systematic ways of building up members in these groups. The BJP has workers dedicated to WhatsApp communications that collect contact files of a localities’ voters and automatically add them to these groups. For the citizens that get added, these can seem like normal community groups. But they’re actually managed and propagated by political parties.
What sort of themes emerge from the data? One of the most prominent themes we’re seeing is how public opinion is discretely influenced by political WhatsApp communications. For example, a huge social and political event unfolded last year in India when Prime Minister Modi inaugurated a new Hindu temple, built over the ruins of a Mosque that was demolished by Hindu political activists. When this inauguration happened the lab analyzed content in WhatsApp groups to look into how people were reacting to and being influenced by those narratives.
What narratives did you see? We joined a range of WhatsApp groups in a small town in the region. In truly organic community groups we saw very little polarizing content or conversations around the Temple. Meanwhile, in political groups and “community groups” managed by political workers, we saw anti-Muslim and pro-Temple narratives being shared alongside polarizing content.
We ultimately saw an effort by political players to promote hateful narratives within community groups and to bring political conversations into the everyday. We concluded that hateful content doesn’t often exist in the truly organic community groups we observed, but that there is an attempt by the political players to spread that kind of polarization elsewhere without actually identifying themselves as political actors. That’s one instance that falls into this broader pattern of how political actors are trying to discreetly influence public opinion on WhatsApp.
You’ve been researching political WhatsApp strategies for a while now. Has anything changed over the years? Political parties’ efforts on WhatsApp have become more sophisticated, automated, and organized. In the past, their WhatsApp groups popped up around elections. Now, they keep these groups alive persistently and keep building on them year-round.
Photo in header: People queue up to vote on Voting Day May 13 in Visakhaptnam Andhra Pradesh for General Elections 2024
Credit: InOldNews
