CGNet Swara — Missed-Call Voice Journalism for Illiterate, Gondi-Speaking Communities in Chhattisgarh, India

CGNet Swara ("voice of Chhattisgarh Net") is a voice-based citizen-journalism platform launched in 2010 by former BBC World Service journalist Shubhranshu Choudhary, working with Microsoft Research India's Bill Thies, to serve rural and tribal communities in Chhattisgarh where under 1% of the population had internet access and many residents are not literate in Hindi or English but speak Gondi and other local languages.

The system works entirely by phone: a caller dials in, hangs up after one ring (a free "missed call"), and the platform calls back, letting the caller record a message about a local issue — a broken hand pump, unpaid wages, a land dispute, a health crisis. A network of roughly 50 volunteer sub-editors, spread across the region, reviews and verifies submissions before they are published for anyone to hear by calling the same number. No smartphone, app, or literacy is required at any step.

A peer-reviewed 2015 study by Microsoft Research and the University of Washington ('Revisiting CGNet Swara and its Impact in Rural India', ACM ICTD 2015), based on 70 interviews and platform logs, documented over 575,000 phone calls, more than 6,900 published stories, and 287 specific grievances resolved as a direct result of a report — including hand pumps repaired and unpaid wages recovered. In one widely cited case, listener-submitted reports surfaced more malaria cases in a single year than the state government had recorded in the previous decade, prompting officials to acknowledge the outbreak.

The platform is still operating as of its most recent (September 2024) annual general meeting and a 2024 Internet Governance Forum session on its work in conflict-affected central India — 14 years after launch, a notable continuity record for a volunteer-dependent model. The 2015 study is honest about limits: most benefits accrue to the subset of callers whose reports get picked up and verified, and impact depends heavily on the unpaid sub-editor network's capacity and reach.

Read the full analysis: https://dl.acm.org/doi/10.1145/2737856.2738026