Sighthound Publishes Its Privacy-and-Evidence Approach to Responsible Vehicle Recognition
Originally published August 26, 2026.
Five Sighthound principles on customer-environment processing, human verification, separate redaction, and agency accountability.
Update: On September 14, the International Association of Chiefs of Police (IACP) and 12 other law enforcement associations published joint model principles for the responsible use of ALPR. They address agency policy, authorized access, documented queries, verification, retention, sharing, security, and accountability. This refresh distinguishes Sighthound's five company principles from the associations' model guidance; it does not suggest IACP endorsement or certification.
LONGWOOD, Fla. — August 26, 2026 — Sighthound Inc., a computer-vision platform company, published its approach to responsible vehicle recognition: a set of architectural principles designed to preserve the public-safety value of automated license plate recognition (ALPR) while answering questions about how ALPR data is governed.
ALPR can give investigators leads about vehicles connected to an investigation, but a read is not proof of a driver's identity or guilt. Communities and policymakers also ask how the data is governed: where it is stored, who can access it, how long it is retained, and when it is shared. Those are questions for agency policy, product configuration, contracts, and applicable law—not a vendor slogan.
Sighthound's position is that many of those answers are architectural, and that vehicle recognition doesn't have to be a choice between public safety and privacy.
"License plate recognition has earned its place in public safety. It closes cases that would otherwise go cold," said Ryan Campbell, CEO of Sighthound. "What people ask about now isn't the camera. It's what happens to the data after it's collected: where it goes, how long it's kept, and who can reach it. Those are fair questions. We built our products so the data stays with the agency, under the agency's rules, and not with us."
Sighthound's approach rests on five principles
1. Process at the edge. Share by choice, not by default.
Installed Sighthound Retriever runs in the customer's environment, and its core recognition pipeline has no required cloud connection. There is no default central or shared national pool of customer reads. Export or sharing is opt-in and customer-configured; the hosted single-image demo is a separate deployment and is not covered by these custody statements.
2. Keep the data with the agency, under the agency's policies.
Installed Retriever stores plate records and media in the customer's environment. That is a data-custody statement, not a claim that the product stores nothing or that current-build retention controls, user roles, or audit exports have all been verified. Agencies should confirm their deployment settings and applicable retention, access, and sharing requirements.
3. Treat a plate read as a lead, and make it easier to verify.
A plate detection does not identify a driver or establish guilt. IACP's joint model principles say a read is an investigative lead that must be evaluated and verified before enforcement action. Retriever returns plate information and vehicle attributes such as make, model, color, and generation; those attributes can inform human review, but Retriever is not independently verified to confirm a plate-to-vehicle match.
4. Let privacy and evidence coexist through redaction.
Sighthound Redactor is a separate file-redaction product for video, images, and audio. Agencies can use redacted working copies to protect unrelated details while preserving relevant context for review. Redactor does not decide what an agency must disclose or establish that a particular release complies with public-records law.
5. Keep a reviewable record of redaction work.
Redactor's 7.3.0 "Acquire Audit Data" archive includes an info.json manifest describing the archive origin and, for each session, a SHA-256 checksum of its audit log alongside video metadata. This is a Redactor file-workflow feature—not a Retriever search audit, per-user edit-level approval trail, or a guarantee of legal chain of custody.
See the Redactor release notes.
These five points describe Sighthound's approach; they are not the IACP's model-principle list, an agency policy, or a certification.
"We read the joint principles from IACP and a dozen other law enforcement associations closely, and we agree with where they land," Campbell added. "Prevent misuse without taking away a tool that finds missing people and solves violent crime. The way to get there is clear policy, controlled access, and accountability, set by the agency. Our job is to build tools that fit that policy, not replace it."
Sighthound offers this approach as a starting point for agencies, policymakers, and community stakeholders working out how vehicle recognition should operate under clearer rules.
To discuss how this approach fits your agency's policies, Talk to sales.
About Sighthound
Sighthound Inc. is a computer-vision platform company. Its products include Sighthound Retriever for ALPR and vehicle recognition, Sighthound Redactor for video, image, and audio redaction, and Sighthound Compute edge AI hardware.
Learn more at sighthound.com.
Media Contact
Roger Dunnavan
Chief Revenue Officer, Sighthound Inc.
