Accountability infrastructure / for surveillance AI.
AI camera systems — automated license plate readers, facial recognition, real-time crime centers, body-worn and doorbell cameras — are being deployed faster than their records can be trusted. The failure that ends contracts and triggers class actions is rarely the detection itself; it is the access log: who queried the system, when, under what stated reason, and who the data was shared with. Today that log is held by the operating agency and the vendor — the parties being held accountable — and in 2025–2026 it has been documented being altered and stripped of fields. Yolo makes that record append-only, independently verifiable, and impossible for the operator or the vendor to edit after the fact — while only hashes ever touch the chain, so no image, plate, or face is anchored.
California ALPR Privacy Act: $2,500 per-person statutory damages, no proof of harm required — the operator's duty to maintain and track access records holds regardless of vendor conduct.
REGULATORY OBLIGATIONS
California ALPR Privacy Act
Active · US (CA)
Civil Code §§ 1798.90.5–1798.90.55
Operators and end-users must track searches of ALPR data, require supervisory approval, and publicly post a usage-and-privacy policy. $2,500 in liquidated damages per violation, no proof of harm required. Recent California appellate rulings hold that deploying ALPR without a compliant public policy is itself actionable.
Illinois BIPA
Active · US (IL)
Biometric Information Privacy Act
Facial-geometry capture requires written consent and a published retention/destruction schedule. $1,000 per negligent and $5,000 per intentional violation, with a private right of action and no proof of harm required. Cumulative settlements exceed $2B.
EU AI Act
Aug 2 2026 / Dec 2 2027*
Art. 12 logging · Annex III biometric ID
Remote biometric identification is high-risk under Annex III. Article 12 requires automatic event logs — timestamps, the reference database consulted, the input data that produced a match, and the verifying personnel — retained at least six months (Art. 26(6)).
FTC Act § 5
Active · US
Biometric surveillance enforcement
The FTC has treated reckless facial-recognition surveillance as an unfair practice, requiring accuracy testing, human oversight, and reasonable safeguards. One national retailer was barred from using the technology for five years.
Washington Driver Privacy Act
Active · US (WA)
SB 6002 — enacted 2026
Restricts government sharing of ALPR data; a vendor's violation is a per se unfair and deceptive act under Washington's Consumer Protection Act.
State ALPR retention & sharing laws
Active · US (multi-state)
CO · KY · ME — 2026
Colorado requires a warrant before federal ALPR sharing; Kentucky caps retention at 90 days; Maine enacted comprehensive ALPR privacy rules. State-official data-sharing audits have found unlawful cross-jurisdiction and federal access.
Evidentiary chain of custody
Active · courts
Rules of evidence
AI-camera-derived evidence must be authenticated with an unbroken chain of custody; edits, gaps, or missing timestamps go to admissibility or weight. Cryptographic tamper-evidence strengthens the showing that a record is in unchanged condition.
EU GDPR
Active · EU
Art. 9 — biometric special category
Biometric data used to uniquely identify a person is a special category requiring a lawful basis and heightened safeguards; processing records must be demonstrable to supervisory authorities.
* EU AI Act Annex III enforcement date: August 2, 2026 (legally operative). EU Digital Omnibus provisional agreement (May 7, 2026) proposes extending to December 2, 2027 — not yet formally enacted. Prepare for the earlier date.
HOW YOLO SATISFIES IT · PRIMITIVE → REQUIREMENT
PRIMITIVE
REQUIREMENT SATISFIED
AUDIT CHAIN
Access/query log integrity
Every query and data-share event logged in an append-only, hash-chained record, Merkle-anchored to Base. A city council, inspector general, attorney general, or court can verify the log has not been altered or had fields removed after sealing — without trusting the agency or the vendor. Directly answers the California duty to track searches of ALPR information.
IDENTITY REGISTRY
Model/version + operator traceability
The detection model version and the querying identity bound to each event are independently verifiable; software changes create documented version events. Supports EU AI Act Article 13 and body-cam chain-of-custody expectations.
DECISIONAL LOGGING
Per-event capture at three tiers
Routine tier for reads and normal queries. Consequential tier for hotlist hits and cross-jurisdiction shares. High-stakes tier for a face match supporting a stop or arrest, a federal disclosure, or evidence sealed for court. Payloads are hashes plus non-identifying metadata only — no image, plate, or face on-chain.
AUDIT CHAIN AND IDENTITY REGISTRY ARE LIVE ON BASE MAINNET TODAY.
HONEST LIMITS
Yolo is tamper-evident, not omission-evident. It proves a logged event was not altered and that fields were not silently removed after sealing. It cannot force an operator to log a query, force a truthful search reason, improve detection accuracy, or make surveillance lawful. Where these systems already exist, Yolo makes their records trustworthy and independently verifiable — that is the claim, and no more.
PRICING · DECISIONAL LOGGING TIERS
ROUTINE
$0.0001 / event
Routine events
CONSEQUENTIAL
$0.01 / event
Consequential events
HIGH-STAKES
$0.10 / event
High-stakes events
VOLUME NOTE
A mid-size department running ~50 ALPR cameras might generate ~5M routine reads/month, ~50K consequential queries, and ~200 high-stakes events — roughly $1,200/month at all three tiers. Access-log-only anchoring (queries and shares, not raw reads) is far cheaper and is the recommended entry configuration. Figures are illustrative.
SCALE
Single agency or private deployer (mall, hospital campus, school): $10K–$75K/year. Mid-size city or county with a real-time crime center: $75K–$400K/year. Statewide fusion center or camera/ALPR vendor OEM integration: $500K–$5M/year.
WHO THIS IS FOR
City councils, police accountability boards, inspectors general, and state attorneys general and their privacy/ALPR audit units; agencies that want to survive procurement and FOIA scrutiny; camera and ALPR vendors that want a credible independence story; and private deployers — retailers, malls, medical centers, schools, HOAs, parking operators — facing statutory-damages class actions.
WHAT THIS REPLACES
Vendor-controlled audit portals that can be edited or stripped of fields. "Trust us" compliance claims that collapse under FOIA and audit. And the litigation cost side: California ALPR at $2,500 per affected person (a single mid-size mall's exposure runs into the hundreds of millions), Illinois BIPA at $1,000–$5,000 per violation (settlements over $2B to date), FTC bans, and wrongful-arrest payouts.
ACTIVATION TRIGGER
A city council or inspector general demands an independent audit after a data-sharing scandal; a state attorney general audit finds unauthorized access and orders verifiable controls; a California ALPR or Illinois BIPA class action is filed against a private deployer; a procurement RFP begins requiring independently-verifiable audit trails; or a prosecutor needs AI-camera evidence to survive a chain-of-custody challenge.
MULTINATIONAL DEPLOYMENT
AI-camera operators face overlapping regimes at once — California and Illinois statutory-damages laws, EU AI Act biometric logging, GDPR, and state-by-state ALPR rules. The same audit chain produces a single independently-verifiable record that satisfies each regime's integrity requirement without re-instrumenting per jurisdiction.
For city councils, police accountability boards, inspectors general, state attorneys general and their privacy/ALPR audit units, agency procurement and compliance leads, and camera/ALPR vendor compliance teams.