Media and platforms · Trust and safety data

Every label knows its policy line. Every face said yes.

Moderation data an audit can trace to your policy, detection pairs the depicted person consented to, and research, verification and content systems with a person at the gate. Reviewed, built and run in the EEA.

The three questions Which line · who reviewed · lawfully

A classifier that flags is easy. Three things make its data usable

Whether the label points to a line of your policy. Whether the person who decided was protected. Whether the pair can lawfully exist.

  1. Which policy line does the label point to?

    HA1 · repeated targeting of a named person · remove

    Every label in the delivery points to one clause of your policy taxonomy, the reviewer's decision sits beside the model's flag, and the two are measured against each other with kappa rather than a percentage, so an audit can trace any decision to the line it was made under.

  2. Who reviewed it, and how were they protected?

    Reviewer r-218 · exposure 4 of 6 · queue paused

    Graphic items arrive blurred and are revealed only by the reviewer's choice, every reveal counts against a per-session cap, and the queue pauses when the cap is reached. The protocol is written before the first item, the reviewer can opt out of a queue without penalty, and the wellness ledger travels with the delivery.

  3. Can the pair lawfully exist?

    Consent c-4471 · depicted person · voice clone vc-2210

    A detection dataset is made of real people. Every real-versus-generated pair carries the consent of the person depicted, a separate consent where a voice is cloned, the generator and its version, and the compression and re-encoding the pair has been through, so a reviewer can trace any pair to the consent of the person depicted.

What YPAI labels, evaluates and builds Six engagements · data and systems · one policy at a time

Start from the policy your platform has to enforce

Three data engagements and three systems, each bought on its own and configured for one policy, one taxonomy and one language set. Four things are settled before the first item, whether it is labelled or built.

Agreed before the first item

  • The policy taxonomy, clause by clause
  • The reviewer protocol and its exposure cap
  • Consent, provenance and residency
  • What the record per item contains
  1. Content moderation training data

    Multilingual, multimodal labels that point to a clause of your policy, reviewed under a wellness protocol.

    Text, image, audio and video items labelled against your policy taxonomy across hate speech, harassment, sexual content, violence, self-harm, fraud and the categories only your platform has, by identity-verified reviewers in the languages your platform runs in. Two reviewers where the protocol asks for it, agreement measured as kappa per clause, and disagreements adjudicated rather than averaged.

  2. Detection pairs and integrity evaluation

    Real and generated recordings of the same consenting person, and a study that says what people and models actually catch.

    Provenance-tagged real-versus-generated pairs across face, full body and voice, with the person's consent for the recording and a separate consent for the clone, the generator and its version, and the compression, re-encoding and cropping the generated side has been through. Then a human-perception study under a written protocol: raters judge pairs blind, agreement is measured, and the result is reported by generator, post-processing and subject rather than as one number.

  3. Video annotation for platforms

    Events, tracking and action labels on the video your platform hosts, with the tracking quality reported.

    Multi-object tracking, video object segmentation, temporal events and action localisation on the video your platform hosts, under a written task protocol with HOTA and IDF1 reported per delivery, so a safety, recommendation or highlight model is trained on labels with their quality attached.

  4. Source desks and verification systems

    A research pipeline that extracts every checkable claim in a draft and anchors it to a source a journalist approves.

    Research and intelligence systems for a newsroom or a verification desk: ingestion of filings, transcripts, datasets and reports, evidence extraction, monitoring, and grounded answers with citations and source traceability. The pipeline proposes the anchor; a person approves it, and the approval is in the record. Built and run in the EEA, on your environment where the archive cannot leave it.

  5. Content and marketing pipelines

    Generation, refresh and outreach in your house voice, with a person at the gate and the model in the EEA.

    Programmatic and research-based content production, content refresh and structured publishing workflows, outreach and enrichment automation, and conversational assistants, integrated with your CMS and your data. A house-voice model fine-tuned on your own material and evaluated against a held-out set, deployed EEA-resident or in your environment, with an approval gate before anything is published.

  6. Moderation agents with an audit trail

    Agent workflows that route, escalate and explain, with human gates and a replay a regulator can read.

    Agent state machines for moderation and trust-and-safety operations: routing by policy clause, escalation to a person at the gates the policy names, retry and fallback, and deterministic replay of every action. An audit-log layer with EU AI Act risk mapping and policy-anchored gating, so the record of why an item was actioned can be produced for a review.

Scope the first queue

Try it yourself: the review queue For the trust and safety lead · one policy · the record per decision

Sit the queue. Decide with the clause. Read what leaves

You are the reviewer. Twelve items come to you as descriptions of what they contain, never the content. The model's flag is shown with its confidence score, a second reviewer has decided blind, and you decide with a clause. Agreement is measured as kappa, and the wellness ledger counts every reveal.

  • The clause is the label.

    Every label points to one line of the policy, and the line is what a reviewer, a model and an audit compare.

  • Kappa, not a percentage.

    Agreement between the reviewer and the model, and between two reviewers, is reported as kappa, which discounts the agreement two raters would reach by chance.

  • A reveal is counted.

    Graphic items are blurred until the reviewer chooses to see them. Every reveal is written to the wellness ledger, and the cap pauses the queue.

Moderation room · Policy P-4 Queue 12 · Decided 0
  1. Queue
  2. Flagged
  3. Decided
  4. Counted

The item

q-13336 Text pl Plain

A comment on a sports result with strong language and no target

Blurred by default. Revealing counts toward your exposure cap.

Model flag SY1 Model confidence 0.62

Your decision

Pick the clause the item breaks, or keep it.

This session

Decided
0
Escalated
0
Exposures
0 / 6

Agreement as kappa

  1. vs the model none
  2. vs reviewer two none

A toy queue of invented items, so the mechanics are visible. Nothing on this page is a real delivery or a real result.

Judge the pair For the detection team · two raters · by post-processing

Two recordings of one person. Which one is generated?

You are rater one. Each pair is a real and a generated toy signal of the same consenting subject, with the generator, the post-processing and the consent on the tag. Rater two has judged blind. Your accuracy, theirs and your agreement are reported with denominators and cut by post-processing, which is what a detection team needs to know where its detector and its people are weakest.

Perception study · Protocol PS-1 Pair 1 / 8

The pair

Pick the side you think is generated. The tag is revealed when you have.

The tag

Pair
none
Subject
none
Generator
none
Post-processing
none
Consent
none
Voice clone consent
none
Rater two
none

This study

Judged
0
You
none
Rater two
none
Agreement
none

Correct by post-processing

  1. None none
  2. H.264 none
  3. Resampled none
  4. Cropped none
  5. Noise none
Toy signals and invented pairs, so the mechanics are visible. Nothing on this page is a real delivery or a real result.

Approve the anchor For the verification desk · every claim · a source a person approves

A draft comes in. Every checkable claim leaves with a source

You are the journalist. An invented draft of eight sentences is on the desk. The pipeline has marked which sentences make a checkable claim, anchored each to the best document in a small corpus with a plain overlap score, and left opinion unsupported. You approve or reject every anchor, and the approval is what leaves.

Source desk · Draft d-0412 Sentences 8 · Checkable 6
  1. Draft
  2. Extracted
  3. Anchored
  4. Approved

The draft

The anchor

Pick a sentence in the draft.

Source
none
Kind
none
Date
none
Overlap
none

Pick a sentence in the draft.

This desk

Checkable
6
Anchored
6 / 6
Approved
none
Rejected
0
Unsupported
2

Anchored Weak anchor No source found Approved Rejected Opinion · not checkable

An invented draft and an invented corpus, so the mechanics are visible. Nothing on this page is a real delivery or a real result.

Who we label for Platform × policy family · configured per programme

Six kinds of platform. What the programme covers for each

A video platform, a marketplace and a newsroom do not fail in the same places. The matrix shows which policy families the programme covers for each kind of platform, in which modalities and languages, and which demo on this page shows the method.

Platform Hate Harassment Violence Sexual Self-harm Fraud Synthetic Modalities Languages Shown in
Hate: In the programme by default Harassment: In the programme by default Violence: In the programme by default Sexual: Its own protocol Self-harm: Its own protocol Fraud: On request Synthetic: In the programme by default Text, Image, Audio, Video EU · Nordic the queue
Hate: In the programme by default Harassment: In the programme by default Violence: In the programme by default Sexual: Its own protocol Self-harm: Its own protocol Fraud: On request Synthetic: On request Text, Image, Audio EU · Nordic · MENA the queue
Hate: Out of scope by default Harassment: Out of scope by default Violence: On request Sexual: Out of scope by default Self-harm: Out of scope by default Fraud: On request Synthetic: In the programme by default Image, Audio, Video EU · Nordic the desk
Hate: On request Harassment: On request Violence: Out of scope by default Sexual: On request Self-harm: Out of scope by default Fraud: In the programme by default Synthetic: In the programme by default Text, Image EU the queue
Hate: In the programme by default Harassment: On request Violence: In the programme by default Sexual: In the programme by default Self-harm: On request Fraud: In the programme by default Synthetic: In the programme by default Text, Image, Video EU · Nordic the study
Hate: In the programme by default Harassment: In the programme by default Violence: On request Sexual: Its own protocol Self-harm: Its own protocol Fraud: On request Synthetic: On request Text, Audio, Video EU · Nordic the queue
  • In the programme by default
  • On request
  • Its own protocol
  • Out of scope by default

Configured per platform and policy. Try the demo named in each row to see its method. Nothing on this page is a real delivery or a real result.

What the delivery contains One row per decision, judgement and anchor · the fields an audit checks

One record. The queue, the study and the desk write to it

Every decision from the queue with its clause, every judgement from the study with its tag, and every anchor the desk approved with its source, in one record with the second rater, the time and whether it counted. Written by what you did above.

Read a row left to right: the kind of entry and its id, what was decided under which clause, source or tag, then the model's flag or score, the second rater's call, the seconds it took, and on the right whether the row counted toward the wellness ledger, the study or the piece.

  1. No entries yet. Decide an item in the queue, judge a pair or approve an anchor above.
Decided
0
Exposures
0
Pairs judged
0
Anchors approved
0

Example record, built from what you ran above

The first queue One policy · one language · one protocol

One policy, one language, one reviewer protocol

A pilot labels one policy in one language under one reviewer protocol, and judges one pair set. The taxonomy, the protocol and the record's fields are fixed before the first item, and the pilot ends with a delivery your safety team can audit and a decision. You judge it against your model.

Fixed before the first item: the policy taxonomy, clause by clause, the reviewer protocol and its exposure cap, consent, provenance and residency, the kappa floor below which a clause is re-briefed, and what the record per item contains.

  1. Policy

    Your taxonomy mapped clause by clause, with the examples each clause turns on.

  2. Protocol

    Blur by default, the exposure cap, rotation and opt-out, written and signed before the first item.

  3. Queue

    One language labelled by identity-verified reviewers, two where the protocol asks, kappa per clause.

  4. Pairs

    Real-versus-generated pairs with consent and provenance, judged blind and reported by post-processing.

  5. Decision

    Continue to the full policy, change the taxonomy, or stop, with the record as the evidence.

The pilot ends with a delivery your safety team can audit.

Scope a pilot

Consent, reviewers and residency GDPR · reviewer wellness · EEA

The controls a review of trust-and-safety data will ask for

A Norwegian company under GDPR. Every real recording carries the depicted person's consent, reviewers are identity-verified and work under a written wellness protocol in our own review console, and the material is stored in Europe by default and processed in the EEA where required.

Jurisdiction
Norway · GDPR-native
Lawful basis
Consent per record · Article 6
Reviewers
Identity-verified · EEA-resident · wellness protocol
Languages
All European languages · Nordic depth · MENA and selected global
Exposure
Blur by default · per-session cap · opt-out
Console
Our own review console · the ledger in the delivery
Audit
A decision record per item · exportable for a review
Provenance
Generator, version and post-processing on every pair
Contracts
Standard DPA terms · SCCs available
Erasure
30-day end-of-contract SLA

Scope a programme A feasibility read · a bounded first queue

Tell us the policy and the platform

We scope against the policy your platform has to enforce, agree the taxonomy, the protocol and the consent model, and test the method on one policy in one language.

  1. Scope
  2. Taxonomy
  3. Protocol
  4. First queue
  5. Delivery

A bounded first queue carries its own acceptance criteria and its own record. You judge it against your model.

Briefs are treated as confidential. We are used to policies that cannot be published and material that cannot leave the EEA.

The brief Tell us the platform, the policy and the languages, and what the model has to catch. We reply with a feasibility read.

What the programme needs (optional)