WhatsApp Became Sri Lanka’s Hidden Election Battleground, Study Finds

Hate speech, disinformation and gendered abuse spread through encrypted party networks during the 2025 local elections, largely out of sight of regulators

WhatsApp became a central but largely unpoliced campaign channel in Sri Lanka’s May 2025 Local Government elections, according to a new study by the Colombo-based research collective Muragala | Centre for Progressive Politics & Policy.

The report, Invisible Trends, describes the messaging app as “a central but opaque campaign infrastructure” that enabled highly coordinated political messaging “that largely escaped public scrutiny and regulatory oversight.” It found hate speech, disinformation, gender-based abuse and breaches of election law circulating in political WhatsApp groups. Most of that content was made on public platforms such as Facebook and then pushed into private groups, where it met little challenge.

The full report is available at muragala.lk.

A rare look inside closed groups

WhatsApp is Sri Lanka’s most widely used social media platform, used by 63% of the population as of late 2024. Because its messages are end-to-end encrypted, what happens in its groups is rarely studied.

To get inside, Muragala trained 30 local activists across all of the country’s electoral districts as observers. Between 10 April and 10 June 2025, they followed 1,743 politically active groups and channels in Sinhala and Tamil, logging suspect posts by hand.

Of 212,134 messages monitored, 2,140 broke election law or amounted to political hate speech, about 1%. But the problem was concentrated: in some groups, up to 30% of messages were violations. Rates dropped sharply where group administrators stepped in.

WhatsApp mostly amplified content rather than created it. Only 4% of violating Sinhala posts and 6% of Tamil posts started on WhatsApp. The rest came mainly from Facebook, shared as links or screenshots, with some from YouTube, X and TikTok.

Hate, falsehoods and gendered abuse

Hate speech and defamation made up the largest share of violations at 34%. Disinformation and technology-facilitated gender-based violence each accounted for 23%, and breaches of the election silent period for 20%.

Nearly half of the hate speech (47%) targeted ethnic and religious communities, mostly Muslims and Tamils. The researchers call this “particularly concerning” given Sri Lanka’s history of ethno-religious conflict. In one example, Muslim and Tamil voters in Colombo were labelled with a derogatory slur for backing the National People’s Power (NPP) mayoral candidate. Another post shared in a WhatsApp group called for the death of President Anura Kumara Dissanayake.

Women in politics were singled out. Prime Minister Harini Amarasuriya, who was not a candidate, was hit by misogynistic and homophobic smears, including a viral video falsely alleging a same-sex relationship. MP Lakmali Hemachandra was mocked over her appearance and mental fitness. Women councillors across parties were portrayed as sexual objects. Intimate images of a female party supporter were leaked without consent, in breach of the Election Commission’s Code of Conduct.

The disinformation followed familiar patterns. Fabricated or exaggerated statistics were most common (32%), followed by manipulated or recycled images (24%), false labels and captions (16%), claims attributed to anonymous “insiders” (16%) and leading rhetorical questions (12%).

Coordinated networks, silent-period breaches

The study documents “cascade” structures, in which content posted in a core group is relayed step by step through layers of linked groups. Some NPP groups were numbered to match Grama Niladhari administrative divisions. The researchers warn this “can also confuse citizens” by making party messaging look state-sanctioned.

On 27 April, 2 May and 5 May 2025, the same false items appeared in groups across several districts within one- to two-hour windows. The authors say this points to a networked, simultaneous messaging system, though the data cannot prove causation.

Campaigning also continued through the 48-hour silent period before polling. More than half of these breaches (56%) came from party-linked accounts, rather than individual candidates.

NPP-aligned groups made up 40% of those monitored and 36% of violating posts, concentrated in a few groups attacking the opposition under the guise of “satire.” Groups linked to the Sri Lanka Podujana Peramuna (SLPP) and United National Party (UNP) produced violations out of proportion to their size: 17% and 16% of violations, from 12% and 11% of groups.

Calls for regulators and Meta to act

The report concludes that WhatsApp’s encryption and closed groups give campaigns “a highly effective, low-cost, and low-risk channel for influencing opinion,” creating echo chambers where “manipulative narratives are reinforced and dissenting voices are rarely present.”

It urges election regulators to bring WhatsApp formally into election monitoring, including enforcement of the silent period, guidance to political parties and collaboration with civil society monitors. It calls on Meta to adopt election-specific rules for WhatsApp groups and Channels, to be transparent about bulk messaging and automation, and to support privacy-respecting oversight during elections.

The authors argue the method itself is part of the answer. Grassroots observers, they write, can document violations “that are invisible to national observers or platform-level monitoring tools,” making them “an indispensable layer of accountability and electoral integrity.”

About the report

Invisible Trends: Monitoring Electoral Violations & Online Harms on WhatsApp through Grassroots Observation was written by Harindra B. Dassanayake, Nipunika Pathirana, Sudaraka Arthanayake, Sajeewa Wijeweera and Manuja Wijesuriya for Muragala | CPPP. It was produced under the Strengthening Resilience Against Disinformation in Sri Lanka project, co-funded by the European Union. The findings cover monitored groups only and are not a representative sample of all WhatsApp use in Sri Lanka.

Read the full report: Invisible Trends on muragala.lk

Tags:

Leave A Comment