SNEK POLICY
Facial Recognition Is a Tool, Not Permission for Permanent Surveillance
Government should not routinely identify and track everyone in public simply because technology makes it possible.
Targeted use, not universal monitoring
Facial recognition can have legitimate targeted investigative uses. That is different from creating a permanent identification system for ordinary people who are not suspected of anything.
Guardrails
Use should have defined purposes, access logs, retention limits, auditing, accuracy requirements, and judicial authorization where appropriate. High-risk matches should be independently confirmed before consequential action is taken.
READ THE FULL POLICY
Why SNEK thinks this
Facial recognition can be useful for identifying a specific suspect or locating a dangerous person, but the same technology can also turn every public camera into a permanent identity checkpoint. SNEK wants a tool that can be used for a defined investigation without normalizing continuous tracking of everyone.
How it would work
Government use should be tied to a specific lawful purpose, documented, auditable, and limited in retention. Broad real-time scanning of ordinary public life should require a much higher legal threshold than comparing an image from a specific investigation against a relevant database. Significant matches should receive human review rather than being treated as conclusive identification by software alone.
Where we’re flexible or still working it out
Rules can differ for border security, access control at highly sensitive facilities, locating a missing person, or investigating a serious crime. The central limit is persistent generalized surveillance of innocent people. Accuracy standards, audit requirements, and warrant rules should become stricter as the surveillance becomes broader or more intrusive.