Political messaging has migrated into private channels faster than any transparency regime has followed. WhatsApp alone reported more than two billion users and has become the primary campaign medium in much of the democratic world — Brazil's elections are run substantially through it, Indian parties maintain distributed volunteer broadcast networks, and U.S. campaigns operate WhatsApp and SMS programs alongside peer-to-texting platforms. The governing fact for accountability is arithmetic: a television ad is one artifact visible to everyone and disclosed to regulators, while a messaging campaign is millions of one-to-one threads that no regulator, journalist or researcher can see. The gap is not a pending policy problem; it is the operating environment.
How do campaigns actually use the channels?
Three techniques dominate the documented playbook. Broadcast lists and channels let one operator push identical messages to thousands of opted-in contacts who receive them as personal messages — the format Brazilian researchers at institutions including the University of São Paulo's research groups have studied most systematically since 2018. Peer-to-peer texting uses human volunteers or paid agents sending scripted but individually sent messages through platforms that keep each interaction inside personal threading, which matters legally: U.S. peer-to-peer texting is largely treated as person-to-person communication rather than automated robocalling under the Telephone Consumer Protection Act, as the TCPA's litigation record established. And group influence operations seed content into existing community groups — neighborhood, religious, professional — where endorsement by a trusted member carries persuasion no advertisement can buy.
What do transparency rules cover — and miss?
The misses are near-total. Campaign-finance disclosure regimes keyed to advertising purchases do not reach volunteer messaging, and even paid texting programs disclose as disbursements without any artifact a voter could inspect. Platform ad libraries — the settlement the industry reached after 2016 — by design contain only paid, public placements. The FCC's robocall rules, reinforced after the January 2024 New Hampshire AI-voice incident, cover automated calls but were built for telephony, not encrypted app threading. The EU's Digital Services Act, the most aggressive transparency regime in force, applies to platforms as intermediaries and reaches private messaging only at its service-review margins — a scope fight Meta contested directly when it challenged DSA obligations extending to WhatsApp's interpersonal features. The result is a clean cleavage: everything public is increasingly legible, everything private is entirely dark.
What has research established about effects?
Less than the alarm suggests, because the method problem mirrors the oversight problem — researchers cannot observe what they cannot sample. Survey-based studies in Brazil and India associate heavy WhatsApp political exposure with polarization and, in some panels, with belief in false claims, but self-report designs cannot establish circulation volume or direction, and leaked datasets are rare. The best-documented intervention is platform-side: WhatsApp's limits on message forwarding — introduced in 2018-2019, capping resends — measurably slowed viral cascades, demonstrating that the friction point is architecture, not content moderation. That finding has aged into the channel's central policy implication: you cannot moderate a billion private threads, but you can change how easily messages replicate.
What are the realistic accountability options?
Three, each with honest limits. Provenance labeling: WhatsApp's metadata indicating a message was forwarded many times, which journalists can instruct audiences to read as a virality signal — the cheapest existing tool. Reporter-side tip lines, which Brazilian fact-checking organizations pioneered at scale, sample the flow without seeing it — a monitoring model, not a transparency model. And spend-side disclosure: requiring campaigns to report messaging-program expenditures and vendor contracts, which would at least expose the operation's size without touching content. None is equivalent to the public ad archive; the choice among them is the choice of which small part of the darkness to illuminate.
What should newsrooms do now?
Treat private-channel claims as a verification beat. The operational standard that has emerged in election coverage: monitor tip lines and party channels for the narratives gaining resend velocity; report the narrative, not the message, with its documented prevalence; and instruct audiences in reading forwarded labels. The sector's uncomfortable conclusion is that the transparency era's victories — ad libraries, spending databases, platform disclosure — secured the public layer of campaigning while the persuasive core moved somewhere the tools do not reach. Covering that layer is now the job's hard part.
For more context, read How to tell a push poll from research that tests harsh messages.
For more context, read televised debate history.
For more context, read political influencer disclosure.