Lochran by Northlight Labs
In plain sight. Off the record.
Hate symbols do their harm in public. The people they target see them every day; the organisations that could act hardly ever do. Lochran closes that gap: sightings captured in the field or ingested from existing queues, and turned into reviewed, attributable intelligence.
In development. First deployments 2026.
The harm is public. The intelligence is wasted.
Every symbol on a wall does two things: it intimidates the people it targets, and it signals where and how hate is organising. Today, only the first one works.
Their targets see them daily
For the communities they're aimed at, these symbols are not abstract: they mark bus stops, school routes, front doors. The harm lands whether or not anyone official ever looks.
Most sightings leave no trace
Crews paint over graffiti every day with no record it existed, and recognising what a photograph shows is specialist knowledge few organisations hold. Sightings go missed, misread, or never logged at all.
Every silent removal wastes a signal
A symbol removed without a record is pattern intelligence lost: which symbols are appearing, where, when, and alongside what. That is the picture Lochran exists to assemble.
From sighting to intelligence
At Lochran's core is a simple, reliable collection pipeline, and everything that enters it leaves as a reviewed record.
1 Capture
Field staff photograph a symbol through the mobile app. Location and time attach automatically, and the sighting joins the triage queue: a dependable collection pipeline from day one, even where nothing was being recorded before. Reporters see the status and, once reviewed, the result.
2 Triage
Incoming images are ranked by urgency, so hate-likely material surfaces first. Where the system is uncertain, it says so: an honest abstain is an answer, not a failure.
3 Review
A person confirms or rejects every identification before it becomes intelligence. Nothing reaches analytics, reports, or partners without a reviewer's decision on the record.
Already holding data?
Email-forwarded photos, bulk uploads, historical backlogs from graffiti and 311 queues: Lochran ingests what you already have and runs it through the same triage and review, so past sightings join the picture too.
Lochran illuminates. It does not adjudicate.
Lochran identifies symbols, not people. Its reference material comes from vetted, published sources, and every identification shows its provenance: which sources say what, and when they were consulted. Where reputable sources disagree about a symbol, Lochran shows the disagreement rather than resolving it.
The criteria for including a source are public. The review history is append-only. The result is intelligence your organisation can cite, challenge, and stand behind, not a private blacklist.
Built with the people who do this work
Designed around existing analyst workflows, not new ones.
Civil-society monitors
Organisations tracking hate incidents get a triage queue that respects analyst judgement, records provenance, and exports evidence in citable form.
Researchers
Symbol-level, aggregate data with sources attached: where and when patterns appear, suitable for validation and publication.
Local authorities
Rapid-removal commitments become operable: the images you already receive, ranked, identified, and reviewed in time to act.
Where Lochran is today
Lochran is in active development by the team behind Pharos, with first deployments planned for 2026. Identification performance is measured against a growing internal benchmark, and we're honest about its limits, including when the right answer is "not sure".
If your organisation sees hate symbols it can't confidently identify, or isn't recording sightings at all, we'd like to hear from you.