Arbiter

See the accounts shaping your feed before they shape your views.

An AI copilot for public discourse · by SimPPL, a U.S. 501(c)(3) nonprofit

SUPPORTED BY GOOGLE · MOZILLA · WIKIMEDIA · FORD · OMIDYAR · MIT · COLUMBIA · UNDP arbiter.simppl.org

What's in your feed shapes what you decide

28%

of Americans trust the mass media, the lowest in Gallup's five-decade trend (Sept 2025).

58%

worldwide worry about what is real and what is fake in online news (Reuters, 2025).

54%

of Americans get news via social media and video, ahead of TV for the first time (Reuters, 2025).

AMERICANS WITH A GREAT DEAL OR FAIR AMOUNT OF TRUST IN MASS MEDIA · GALLUP

SRC · news.gallup.com/poll/695762 · Reuters Institute Digital News Report 2025 ARBITER · SIMPPL

Manipulation reaches the decisions that matter

Jobs, health, elections: real, documented cases, sourced on every card.

BBC News: Cambodia scams, lured and trapped into slavery in South East Asia — Jobs · trafficking
Fake online job ads lure workers into South East Asia's forced-labour scam compounds.
Src · BBC News ↗

The Guardian: Samoa measles outbreak, WHO blames anti-vaccine scare as death toll hits 39 — Health · vaccines
Online anti-vaccine fear dropped Samoa's vaccination rate below 30%; a measles outbreak killed dozens of children.
Src · The Guardian ↗

Al Jazeera: Romania's top court annuls results of presidential election's first round — Elections · Romania
A court annulled Romania's presidential election over a coordinated TikTok operation; the evidence later vanished.
Src · Al Jazeera ↗

Three documented cases · full source on each cardARBITER · SIMPPL

Campaigns cross platforms. Moderation stops at each platform's edge.

On Twitter, true stories took about six times as long as false ones to reach 1,500 people (MIT study, 2018). By the time a claim is checked, the campaign that planted it has moved on.

We rely on journalists and fact-checkers to catch this, but they've lost their view: CrowdTangle shut down in 2024, and paid access now costs thousands a month. And each platform moderates only its own feed, so a campaign across X, YouTube, Telegram, and WhatsApp shows each watcher a quarter of the picture.

OnPassive, an "AI-powered" tech scheme, recruited across social platforms at once. The SEC charged it with fraudulently raising over $108 million from more than 800,000 investors. Each platform saw only its own slice.

SRC · MIT (Vosoughi et al., 2018) · SEC v. Mufareh / OnPassive, LR-25809 ARBITER · SIMPPL

Arbiter reads public conversation across platforms

01 · COLLECT

Public data, collected within platform rules

We help collect public social data from sources within platform terms of service, including platform-granted APIs, across X, YouTube, Reddit, Bluesky, Instagram, Facebook, TikTok, LinkedIn, Telegram, and 4chan, plus 20,000+ newspapers and Wikidata.

02 · AGGREGATE

Themes, actors, networks

Posts cluster into themes and sub-themes. Each theme surfaces the accounts moving it, with per-actor dossiers and influence maps.

03 · INVESTIGATE

Agentic investigations

Deep research agents answer your questions across the corpus and cite every post they draw on. You keep the judgment.

3.7M+

posts collected across investigations

20B

interactions a month across platforms

100+

organizations with journalists signed up

countries with newsroom and civil-society partners

SRC · arbiter.simppl.org live counters (May 2026) · simppl.org/about ARBITER · SIMPPL

Platform enforcement lines up with what Arbiter flagged

X 2023

Our 2023 reporting prompted an internal investigation at X into pro-Russian bot networks.

Investigation opened

Meta Q1 2024

We shared findings with Meta's team; their public Q1 2024 Adversarial Threat Report details the takedown of a Bangladesh network: 50 accounts and 98 Pages with 3.4M followers.

Network removed

YouTube 2025

Channels we flagged for AI-generated propaganda were terminated by YouTube independently. We had filed no report; enforcement confirmed the detection.

Channels terminated

SRC · transparency.meta.com/sr/Q1-2024-Adversarial-threat-report ARBITER · SIMPPL

Trace conversations across platforms

Pick a topic or a set of accounts and a date window, and Arbiter compiles a month of cross-platform conversations in 20 to 40 minutes. We've built reports on over 30 issues this way, from job scams and human trafficking to the manosphere influencers reaching younger audiences.

XYouTubeRedditBlueskyInstagramFacebookTikTokLinkedInTelegram4chan20k+ newsWikidata

Every study keeps its sources attached: each number traces back to the posts behind it.

arbiter.simppl.org/studies

SRC · arbiter.simppl.org/studies · product screenshot ARBITER · SIMPPL

arbiter.simppl.org · topic modeling

Themes surface the actors moving them

Posts cluster into themes, themes into sub-themes, and each cluster names the accounts driving it. The graph turns a feed into a map: who amplifies whom, in which language, on which platform.

Reports built this way sit behind both platform outcomes on the track-record slide: X's internal investigation and Meta's Bangladesh takedown.

SRC · product screenshot ARBITER · SIMPPL

Ask the agent a real question

We put this to a live study of the manufactured attention around Ibrahim Traoré, the officer who seized power in Burkina Faso. The agent read all 191 posts and ranked the accounts driving the conversation.

You interrogate a corpus too large to read in plain language and get an answer you can act on. The engagement traces to a handful of accounts spiking in single bursts rather than sustained volume, the shape of a coordinated push, not a real audience.

Every row ties back to the posts behind it. A question like this costs 5 credits and takes under a minute.

traoré study · 191 posts · 3 platformsMAY 26 – JUN 2, 2026

Which accounts drive the most engagement on Ibrahim Traore content in this window, and where do their outbound links point?

ANALYZING PATTERNS…

Engagement is driven by a small set of accounts, mostly via high-performing posts that don't include outbound links: native video and text rather than link-sharing.

Account	Posts	Interactions	Outbound links
zoomafrika1	1	56,860	none detected
SahelAlerte	1	29,721	none detected
WelcomeTheGulag	1	23,175	none detected
kelevitch	1	22,426	none detected
WithoutHistory	1	10,499	none detected
kwadwosheldon	1	8,725	youtube.com

"Driver" status here is concentrated in single spikes rather than sustained volume.

Which accounts are amplifying this narrative the mostHow did this discussion spike and evolve over time

SRC · live agent answer, Jul 8 2026 · question, table, and takeaway verbatim, prose condensed (research/09) ARBITER · SIMPPL

One tool, reading across every feed

Search the posts on every platform, aggregate the ones that move together, then investigate the accounts behind them.

01 · SEARCH STACK

YouTube

02 · AGGREGATE STACK

Coordinated praise

44 posts · 3 platforms

Recycled talking points

31 posts · 2 platforms

03 · INVESTIGATE STACK

one hub drives the rest

SRC · Search · Aggregate · Investigate stack · arbiter.simppl.orgARBITER · SIMPPL

Under the hood: three problems we solve through rigorous science

01 · SEARCH

Find the posts you meant, not the words you typed

The language that carries a story is coded, multilingual, and shifts week to week. The hard part is retrieving what someone intended even when the obvious keywords would miss it.

02 · AGGREGATE

Map who drives which narrative, across languages

A feed of millions of posts hides its own structure. Turning it into a navigable map of narratives and the accounts moving each one, resolved to real entities, is more than clustering.

03 · INVESTIGATE

Interrogate millions of posts on demand

Ask any question of the corpus and an agent writes and runs the analysis, returning a cited answer. Surfacing custom insight on demand is the value; grading each answer against baselines is what makes it safe to rely on.

Most listening tools answer with RAG over whatever their feeds already collected. Arbiter plans the collection itself and surfaces conversations you did not know to ask for. The next slides show the method; the appendix lists every model.

SRC · SimPPL methods talk, NYU CSMAP (Mar 2026) ARBITER · SIMPPL

Search: your question finds posts you never typed

>Climate skepticism and pushback on glacier-melt reporting

01 · PLAN

Your question expands into a typed plan: the actors, places, and domains it implies, from a daily-refreshed knowledge graph.

            ORGANIZATIONS · Danish Met Institute · Columbia Climate School

            LOCATIONS · Antarctica · Greenland · Andes

            DOMAINS · sepp.org (skeptic) · nature.com

            PLATFORMS · X 68 · YouTube 46 · Bluesky 16

02 · RETRIEVE & RERANK

Keyword and meaning search, fused, then reranked to what actually answers. One threshold drops 60 to 80% of the noise.

10,231

collected

→

138

relevant

03 · THE PAYOFF

No one labels their own post skeptical. Search the word skeptic and you find almost nothing. The pushback itself, "the models always run hot," "ice has advanced before," shares none of your words, and Arbiter surfaces it by meaning.

81% → 99%

relevant results in the top 200, after reranking

SRC · SimPPL retrieval pipeline · entity prefilter cuts rerank GPU ~99% · climate-skepticism case, arbiter.simppl.org ARBITER · SIMPPL

Arbiter maps the conversation beyond your search terms

A keyword tool returns matches. In a separate April 2026 window we asked one question about Ibrahim Traoré, collected 22,719 posts, and kept the 19,532 relevant ones. Arbiter aggregates them into a map you can walk, concept to sub-concept to post, including conversations the query never mentioned. click any row below

Concepts 9 of 51 shown

Governance Challenges in West Africa1253

Cohesive National Development883

Sovereignty and Cultural Resilience395

Strategic Governance and Economic Transformation370

Political Discourse and Rights310

Political Authority and Resistance Dynamics264

Cultural and Political Empowerment258

International Relations and Reconciliation60

Judicial Inquiry Narratives24

Sub-concepts 7 of 41

National Unity and Progress719

Bobo-Dioulasso Event Highlights68

Historical Prison Visit Impact58

Western Criticism Response12

Unfiltered Interview Insights12

Leadership and Patriotism8 ▾

April Ministerial Resolutions6

Posts 3 of 8 · Leadership and Patriotism

TrueMamle YOUTUBE

Captain Ibrahim Traore Speech 1st Of April : Revolutionary Spirit Of Burkina Faso (English)

504 interactions

Bridge24 YOUTUBE

Interview of Captain Ibrahim Traoré | President of Burkina-Faso | (Part1)

9 interactions

FLPLOUZANE TV YOUTUBE

News : President for life' Inside Traoré's Burkina Faso (UK, 13 april of 2026)

2 interactions

The YouTube tab of this study holds 51 concepts, 41 sub-concepts, and 3,817 posts from 1,115 users; this slide walks one path. The rest open in the live study.

SRC · Ibrahim Traoré study, Apr 1–30 2026 window, YouTube concept web · captured Jul 2026 ARBITER · SIMPPL

Aggregate: geometry stops theme labels from repeating

Say a search comes back with 100,000 posts. Point an LLM at them naively and the bill explodes while the labels collapse into duplicates: 28 of 53 clusters in one corpus came back with the same label.

We group the posts by meaning first, in a geometric space, before any LLM reads them. It is token-efficient where frontier-model approaches are not, a first-of-its-kind method we were invited to present at Stanford.

The LLM then labels whole clusters, so cost scales with a few dozen clusters, not a hundred thousand posts. Any label that lands too close to an existing one is thrown out.

The LLM provides the vocabulary. The geometry enforces distinctness.

            NAIVE   Climate Change · Debate · Discussion · Discourse (all the same)

            ARBITER Greenland sovereignty & Arctic ice · the polar vortex · China–Pacific climate diplomacy · permafrost & policy

"AVOID PRIOR LABELS"0.509

POST-HOC DEDUP0.443

TWO-PHASE EXTRACT0.438

OURS · HYBRID0.231

The climate query on arbiter.simppl.org/features: naive labeling collapses to one theme; ours keeps them distinct. (Bars: average pairwise label similarity, shorter is better.)

The same embedding space powers entity extraction: nine entity types, resolved to Wikidata, name variants merged automatically. Each actor gets sentiment, emotion, and a clickbait score on a five-point scale.

SRC · SimPPL labeling benchmark, silhouette 0.71 vs 0.55 K-Means · arbiter.simppl.org ARBITER · SIMPPL

Deception signals Arbiter reads in every corpus

Sentiment uniformity

Discourse that agrees too much

The Traoré PR campaign read 100% positive, 100% joy across coordinated pages. Organic discourse rarely does.

Clickbait score

Bait at 5, camouflage at 1

The top jailbreak seller scores 4.2. The Traoré operation scores 1, polished to pass as organic praise.

Amplification anomaly

Reach with no audience trail

One crypto casino drew 36.9M interactions from four identical posts; one Traoré post drew 88.8K alone.

Entity mention share

Who benefits from the story

1win led the betting corpus at 90 mentions while mirror domains kept the banned brand alive.

Impersonation

Borrowed legitimacy

Entity resolution surfaced one page posing as 16 real manufacturers, Indofood to Yamaha, recruiting into scam compounds.

Sponsorship disclosure

Paid voice posing as organic

Influencer dossiers separate disclosed from undisclosed promotion and flag paid-promotion patterns that break India's securities-disclosure rules.

Arbiter sits as an analytical layer over social platforms and online sources. These signals add up to tactics: manufactured consensus, coordinated amplification, borrowed legitimacy, undisclosed promotion. We test the classifiers in non-Latin scripts too, from Hindi to Cyrillic, so manipulation in those languages isn't missed.

SRC · public Arbiter studies + partner case reports · arbiter.simppl.org/studies ARBITER · SIMPPL

Investigate: the agent digs, and cites every post

Ask a question and the agent runs the investigation. It picks from 30+ analysis functions, then writes and runs code against the corpus in a sandbox.

Every answer arrives as an artifact you can verify: the code it ran and the source posts it cites.

Every answer is also graded automatically, and the error patterns feed continuous improvement, so each fix reaches every future query.

You keep the judgment. The agent does the digging.

            > What's driving the climate skepticism debates?
          

ARBITER INTELLIGENCE

KEY NARRATIVE

"Climate change is a hoax, clean energy is a scam": 45% of all interactions

TOP SKEPTIC ACCOUNTS BY REACH

@BGatesIsaPyscho 2.4M · Sen. Matt Canavan 1.1M · @MatrixMysteries 859K

DECEPTION SIGNAL

The loudest account, an anonymous conspiracy page, scores 4.2 / 5 on clickbait and emotional manipulation

Every line links back to the source posts. From the "Narratives regarding Climate change" study on arbiter.simppl.org.

SRC · Arbiter Social Intelligence Agent · arbiter.simppl.org/features ARBITER · SIMPPL

Accessing social data isn't the hardest part

Collecting posts is the easy part. The hard part, and what we build, is turning them into evidence you can trust. Five problems make it hard:

Vendor sampling

Every feed is a sample, and the vendor's sampling choices ride into every count computed downstream.

Continuous analysis

Pipelines have to keep running over new posts and their images, video, and audio, long after the first export.

Cross-platform patterns

Campaigns move across platforms, languages, and countries. The analysis follows them or misses the pattern.

Evidence of harm

Harm gets labeled, flagged, and documented to a standard that platforms and regulators will act on.

Near real time

All of it lands faster than the harm reaches a larger audience, or the finding arrives too late to matter.

SRC · SimPPL data-pipeline practice · arbiter.simppl.org ARBITER · SIMPPL

Case file 01 Burkina Faso · AI-generated PR campaign Jun 30 – Aug 30 2025 · X · YouTube · Reddit

01Detect

A cluster of accounts pushes a flawless president

Coordinated pages promote Ibrahim Traoré with sentiment reading 100% positive, 100% joy. Organic discourse rarely reads this way.

02Measure

Engagement without an audience trail

One account lands 88.8K interactions from a single post (views, comments, and likes combined). Its clickbait score is a clean 1 of 5. The campaign is built to look like ordinary fan praise, and to a casual reader it does. Arbiter flags it anyway: organic audiences disagree with each other, and this one never does.

03Identify

The videos are AI-generated slop

Fake factories, fake railways, fake tributes, drawing millions of views before any platform labels them.

ibrahim traoré · key narrative

Traore panel: 100% positive sentiment, 100% joy, Info Pulse top post

POSITIVE100%

JOY100%

Uniformly green panels are the anomaly that starts the investigation.

SRC · public Arbiter study panels · arbiter.simppl.org/studies ARBITER · SIMPPL

Case file 01 The campaign's claims, checked Verdicts · enforcement · 2026 follow-up

The claim: new factories, railways, and mega-projects across Burkina Faso. The flagship videos celebrating them are synthetic.

AI-generated

The claim: an organic pan-African support wave. Uniform 100%-positive sentiment and single-post interaction spikes match coordination patterns, and platforms later labeled videos as altered.

Coordination pattern

The claim: harmless fan content. YouTube terminated the flagged channels on its own; we had filed no report.

Enforced against

Early detection · validated by YouTube

One year on (May 26 to Jun 2, 2026), a rebuilt network of 30 channels cut from one naming template works the same beat, now with tip jars, memberships, and merch links attached. One of the 2025 actors is back under the same name.

youtube.com · flagged channel

"This video is no longer available because the YouTube account associated with this video has been terminated."

SRC · study panels + subsequent platform enforcement ARBITER · SIMPPL

Case file 02 Fake overseas jobs, Indonesia and the Philippines Dec 2025 – Apr 2026 · Facebook · Migrasia

Two studies built with Migrasia read 2,251 recruitment posts from 1,444 accounts. Most scam posts work hard to look legal: 489 Indonesian posts open with words like "resmi" (official) and "no upfront fee", and 183 Filipino posts print a recruitment licence number.

The contact details give the networks away: one shortened link posted by 21 regional job pages under different names; six profiles that all route to a single WhatsApp number; in the Philippines, 34 shared phone numbers and emails tie the pages into 14 networks.

One agency kept posting vacancies under licence DMW-011-LB-051822-R after that licence had expired, so the licence number itself was part of the sales pitch.

14 recruiting networks mapped

LEGITIMACY SCRIPT489

TESTIMONIALS229

URGENCY167

PHONE NUMBER154

WHATSAPP124

Recruitment tactics by post count, Indonesian corpus (1,492 posts); bars scaled to the leader.

A quarter of the Filipino corpus is the aftermath: airport interdictions, arrests for illegal recruitment, repatriations of trafficking victims.

SRC · two public Arbiter × Migrasia reports, linked on the partners slide · arbiter.simppl.org/studies ARBITER · SIMPPL

Case file 03 Manosphere narratives, packaged for young audiences Jan 1 – Feb 28 2026 · X · 3,361 posts

Two months of high-engagement posts drew 127.3M interactions. The content is sold as self-improvement: "looksmaxxing" routines and testosterone regimens, with verified health information pushed aside.

Engagement runs on a star system: the top 3 accounts take 14.6% of all engagement, while the highest-volume cluster (1.8K posts) registers 0.0%. Culture-war framing carries 57% of interactions; belonging and dating discourse carries 56.8%.

The report maps outbound audience redirection: the links actively shared to move audiences off-platform, and the content driving them there.

3 accounts · 14.6% of engagement

CULTURE WAR57.0%

BELONGING56.8%

ANALYTICAL TAKES9.1%

Share of total interactions carried by each theme (themes overlap, so shares exceed 100% combined).

Inside the conversation, a few accounts like @DrewPavlou and @MikhailaFuller carry it, each turning one or two posts into millions of interactions. The recurring claims run from testosterone decline to dating advice, wrapped in culture-war and media-accountability framing; Netflix's documentary was one entry point.

SRC · Understanding the Manosphere report (Arbiter, Mar 2026) · arbiter.simppl.org/studies ARBITER · SIMPPL

Case file 04 Maryland's mail-in ballots: record vs feed May 1 – 19 2026 · X · YouTube · Facebook · Bluesky

What officially happened

A vendor printing error sent some voters the wrong party's primary ballot. Unable to identify who was affected, the state re-sent replacements to everyone mailed a ballot before May 14, covering 500,000+ requests. Originals were voided; every return envelope carries a unique identifier, so officials state there is no risk of duplicate voting.

state board explanation, cross-platform

Maryland State Board of Elections explanation cited across three platforms

The official explanation, cited in the study.

What the feed made of it

The claim: 500,000 fake or illegal ballots revealed. False per the state's election administrator and independent fact-checkers.

Rated false

The claim: a million ballots now circulating. Originals plus replacements, recast as duplicates; the study's claims review found the leap unsubstantiated.

Escalation

One templated line, 50+ accounts. The same wording burst across platforms within hours, overwhelmingly from accounts that are commentary rather than news outlets.

Template bursts

The stakes were legislative: the study's top post (1.32M interactions) demanded the Senate attach the SAVE America Act, which would require in-person documentary proof of citizenship to register and would replace automatic mail-ballot delivery with application-only mail voting. The false ballot claims circulated as the Senate weighed it; the act passed the House 218–213 and failed in the Senate.

SRC · 1,485-post study, saved claims + coordination runs · elections.maryland.gov · FactCheck.org · AP · congress.gov (H.R. 7296) ARBITER · SIMPPL

Technical advances in cross-platform discourse analysis

We tested our own pipeline against hand-labeled data before trusting it. Two results changed how we build.

A plain entity filter beat existing neural rerankers

Fusion scores are not comparable across queries, so no fixed cutoff works. We calibrated the reranker to emit an absolute relevance score, then found a deterministic entity prefilter matched it on most candidates for almost no compute.

on-topic results, top 200 81% → 99%

rerank compute per case study ~$50 about 99% less GPU than reranking every candidate

Measured on replayed case studies against hand-labeled posts.

Geometry keeps theme labels distinct, not the LLM

An LLM labeling clusters one at a time repeats itself. So the LLM only proposes vocabulary; the embedding geometry discards any label within 0.7 cosine of an existing one. Nine entity types resolve to Wikidata, name variants merged automatically.

unique cluster labels naive: 28 of 53 dupes → 98% unique

clustering quality, silhouette 0.71 vs 0.55 K-Means

Silhouette from our labeling benchmark. Cost scales with a few dozen clusters, not 100k posts.

SRC · SimPPL retrieval and labeling pipeline · code: packages/retriever, packages/analysis ARBITER · SIMPPL

Scaling agentic investigations to millions of posts

The agent writes and runs TypeScript in a sandbox against the corpus, picking from 30+ analysis functions, capped at eight steps. Every answer returns the code it ran and the posts it cites, and any run replays deterministically.

3.7M+

posts across investigations

20B

interactions a month across platforms

9 + news

platforms, multilingual, entity-resolved

540

agent runs hand-graded across 9 studies

We grade the agent's code, not just its prose

An agent that chooses tools and writes code can be confidently wrong in quiet ways. An LLM judge scores each answer's code against a reference implementation we wrote, and error patterns feed an offline GEPA prompt-evolution pass our engineers review.

analysis functions the agent composes 30+

every answer replays from its code deterministic

Grading executes the agent's actual code against a known-good baseline, so an answer is only accepted when its code reproduces the reference result.

SRC · code: run-case-study-agent.ts (8-step TS sandbox), packages/agent-eval, tooling/scripts/replay-* ARBITER · SIMPPL

Where Arbiter fits in your workflow

01 · SCOPE

You bring the question

A tip, a beat, a policy window. We scope the corpus, platforms, and dates together.

02 · RUN

The pipeline does the sweep

Retrieval, themes, and actor dossiers land in a case-study library. Your analysts question it through the agent, with a citation on every answer.

03 · PUBLISH

You keep the evidence

The findings are yours to publish. Evidence stays attached, so editors and reviewers can audit every number.

Monitoring and OSINT desks

One scoped corpus replaces manual sweeps across four platform search bars.

Fact-checking desks

In Mongolia, NEST read a defamation bill's direction in the data 19 days before an eyewitness confirmed it.

Trust & safety teams

Coordinated campaigns surface before they reach your abuse reports.

Narrative-intelligence products

Our scored corpora and dossiers slot in as an analytical layer under your platform.

Model stages are commodity parts; the retrieval planning, the thresholds, and the grading harness are ours. When a better model appears, we swap the part and your workflow does not change.

SRC · pilot scoping: team@simppl.org · platform: arbiter.simppl.org ARBITER · SIMPPL

Newsrooms, civil society, and agencies in 21 countries

100+ organizations already use Arbiter, on five continents.

Rappler · PHJagran · INGothamist/NYPR · USFactchequeado · LATAMAfrica Uncensored · KENEST · MNMigrasia · HK/ID/PHDW Akademie · dpa · DEUNDP · UN

Newsrooms, fact-checkers, UN trust-and-safety teams, and U.S. state election officials. YouTube grants us 1.2M API calls a day. Stanford picked Arbiter for this year's platform workshop; Meta, YouTube, and TikTok held that slot in past years.

SRC · arbiter.simppl.org partner strip · simppl.org/aboutARBITER · SIMPPL

Built by researchers from Twitter, Adobe, and Slack

Swapneel Mehta

Founder & President

Former postdoc in AI safety at MIT and Boston University. Board member at Integrity Institute. Built AI systems at Twitter, Adobe, and Slack. Google Research Innovator.

Dhara Mungra

Co-founder · Engineering lead

Led audience-analytics ML as Data Scientist III at Bombora. M.S. in Data Science at NYU. Fellow with the Center for AI and Digital Policy and at the University of Mannheim.

Six full-time engineers

Utkarsh Verma · Atmik Shetty · Dev Bhut and team

System design and infrastructure, NLP and deep research agents, platform backend and APIs.

Why trust us

Credibility

14 peer-reviewed papers (NeurIPS, ICML, AAAI, ICWSM). Awards from Google, Mozilla, Wikimedia, Ford, and Omidyar. Bootstrapped as a nonprofit since 2021.

SRC · simppl.org/team · simppl.org/research ARBITER · SIMPPL

What Arbiter does, and what stays human

The agent

Collects public social data within platform terms of service, clusters themes, maps actors and networks, and answers questions with a citation for every claim.

The human

Draws the conclusions. Journalists, researchers, and trust & safety teams keep the judgment; Arbiter keeps the evidence attached.

Scope

We fact-source: the system surfaces claims and groups related ones, so one verified fact-check scales to thousands of posts. We spent a year building this with Wikimedia.

The case files in this deck are public studies from 2025 and 2026 windows. A study scoped to your issue tracks your actors, in your languages, on your timeline.

SRC · arbiter.simppl.org/studies · methods grounded in peer-reviewed research ARBITER · SIMPPL

Come in at a tier, or anchor the whole plan

Access · $500K a year

Train journalists in 12+ countries

New cohorts on our Kenya model (10 journalists with Africa Uncensored and DW Akademie), and the platform stays free for independent journalists and fact-checkers.

Languages · $250K a year

Reliable AI in partner languages

Mongolian, Swahili, Bahasa, and Hindi first, with the multilingual benchmark published openly.

Evidence · $250K a year

Double the public case library

From 30+ studies to 60+, audience reach instrumented, and the AI held to public calibrated evaluations.

Each activity is an annual tier you can lead on its own. Fund all three for a year: $1M. The full 24-month proposal is $1.8M. A scoped pilot on your own issue starts at $25K–$150K; anchoring a new country vertical runs $1.5M–$3M.

Grants subsidize free access; paid subscriptions with larger newsrooms and civil-society clients grow each year. first recurring revenue in 2025 at $15K a quarter. Since 2021 we have brought ~$450K to SimPPL from grants raised jointly with collaborators.

SRC · 24-month proposal and business plan on request · founders@simppl.org · SimPPL Annual Report 2025ARBITER · SIMPPL

Try it now.

Request a demo: write to team@simppl.org
and we follow up within the week.

A pilot is one case study scoped to your issue: we set the date window together, run the investigation with you, and you keep the evidence and the sources. Our grants subsidize free access for independent journalists and fact-checkers, and paid subscriptions with larger newsrooms and civil society clients generate growing revenue each year to support our moonshot projects and team members sustainably. Free investigation reports land in your inbox via simppl.org/newsletter. The platform is live at arbiter.simppl.org.

SIMPPL · U.S. 501(C)(3) · BOOTSTRAPPED SINCE 2021 ARBITER · SIMPPL

Appendix: models in the pipeline, and why

Query planning

GPT-5-nano

Constrained decoding for typed facet decomposition

Post embeddings

Voyage 3.5-lite (1024-dim)

Cost and quality tradeoff at corpus scale

Neural reranking

Voyage rerank-2.5

Absolute relevance scores that hold across topics

Entity extraction

gpt-oss-120b via Groq

Fast open-weights inference; 20b fallback

Theme labels

GPT-4o-mini

Best quality at lowest cost in our labeling benchmarks (GPU theme service)

Entity narratives

GPT-5-nano

Lightweight generation with a rule-based fallback

Agent task LLM

GPT-5.2

Tool selection and code generation

Agent judge

Model-agnostic (OpenRouter)

Six-dimension rubric; Claude Sonnet 4.6 judged the 540-run campaign

Prompt optimization

GEPA + Claude Sonnet 4.6

Prompt evolution via natural-language reflection, run offline.

APPENDIX · MODELS AS OF JULY 2026 · PARTS SWAP AS BENCHMARKS MOVE ARBITER · SIMPPL