SimPPL

Rebuilding trust in online information.

Everyone deserves reliable information to make decisions.

We build public-interest AI tools in service of human agency online. Our work defends civic discourse that shapes elections, healthcare, and public policy.

75+
journalists using our platform
2B+
monthly reach of content we audit
8
countries
Q1 2024 · Bangladesh
National-news feature →
Meta safety action
2025 · Burkina Faso
AI slop, flagged
before YouTube labeled it
What we ship

Unifying the platform, the mechanism, and the community to rebuild trust in online information.

Arbiter is the platform that shows the public what is shaping their feeds. Sakhi operationalises the same trust mechanisms to deliver reliable information for antenatal care. NextGenAI builds the community of engineers who will ship the next wave of responsible AI.

arbiter.simppl.org/investigations/synthetic-media-q3-2025
NarrativesNetworksLLM agentsCross-platformQ3 2025 · 6 platforms
Flagged narrative · inauthentic amplification

Coordinated amplification of a civic narrative by inauthentic accounts

Authenticity score
low
Cluster · 14 nodes
Top microinfluencers
  • @civic_voice_officialYouTube · 2.4M subs
    0.84
  • @truth_uncensoredX · 480K followers
    0.71
  • @news_aggregator_24Telegram · 6 channels
    0.68
Citation chain
inauthentic account seeds claim+ 4 reposts in 9 minutes+ cross-platform amplification+ AI-voiced derivative on YouTube
Platform · Arbiter

Arbiter shows the public what is shaping their feeds.

Arbiter is a public observatory for digital narratives. Journalists, regulators, and civil-society researchers use it to trace stories across the popular mainstream social platforms, where content reaches over two billion people each month. Built with platform-approved data and local partners worldwide.

Sakhi field visit, Jalgaon, January 2025.
Mechanism · Sakhi

Sakhi operationalises the same trust mechanisms, for antenatal care.

SimPPL's mission is to surface reliable information for people to trust, and Arbiter has always been the north star for the lab. With Sakhi we operationalise the same mechanisms to deliver reliable information to individuals, and measure whether it can change antenatal care practices. A multilingual WhatsApp chatbot in Marathi, Hindi, and Bangla, built with frontline health workers and clinicians, evaluated inside two field RCTs in India and Bangladesh.

Atlantic Dialogues, Morocco, 2024.
Community · NextGenAI

NextGenAI trains the community of engineers who will ship the next wave of responsible AI.

A Responsible AI program for undergraduate teams in India, supported by Google and Mozilla across two fellowship tracks. 82+ students have come through the program. Sakhi spun out of NextGenAI.

Since 2017

How we got here. Six moments that shaped what SimPPL is now.

  1. 2017
    01

    DJ Unicode founded at DJ Sanghvi College, Mumbai.

  2. 2021
    02

    SimPPL launched as a research collective.

  3. 2023
    03

    Mozilla Responsible Computing Challenge inaugural India award. WikiCred funds Arbiter.

  4. 2024
    04

    Incorporated as a U.S. 501(c)(3). Featured in Rest of World ahead of Meta Bangladesh CIB takedown.

  5. 2025
    05

    Sakhi Jalgaon endline. Cohere AI multilingual reproductive-health benchmark.

  6. 2026
    06

    Selected for the Fast Forward Tech Nonprofit Accelerator.

Areas of research

Social, economic, and technological drivers of digital trust.

We work where the three drivers meet: behavior, incentives, and machinery. Each area below names what we are investigating and what we have published recently.

Social drivers

How communities form, share, and update beliefs online. User controls and platform governance. Cross-platform narrative coordination.

  • User controls and engagement on Bluesky (n = 110K blocking events)
  • Coordinated cross-platform narrative tracing
Economic drivers

How attention markets shape what reaches whom. Source credibility under scale; pink-slime websites and AI-generated news.

  • Domain credibility from machine-harvestable indicators (regression vs LLM)
  • Market-design interventions for safer agentic AI
Technological drivers

Multilingual evaluation for low-resource languages. Reproductive-health LLM benchmarks. Human-in-the-loop review of AI medical advice.

  • Multilingual reproductive-health benchmark (with Cohere AI)
  • Health-Eval clinician review framework
Where SimPPL has shown up

Recent presentations.

Follow on LinkedIn
Talk to us

Funder, newsroom, regulator, public-health partner.

We respond to every email.