Your phone will automatically search for the nearest available signal and connect to it. FalcoNet targets exactly that behavior. The NYPD is seeking public comment on a proposed three-year sole-source contract with Cognyte Technology Inc. The contract is not yet finalized.
What FalcoNet actually does
The device mimics a legitimate cell tower, capturing identifiers from phones that connect to each other before police can narrow in on a target’s location.
FalcoNet impersonates a real cell tower, and close phones connect to it as a result. The system then captures device identifiers and helps investigators determine the relative signal strength and general direction to a device. It does not produce GPS-level location data.
This ability is not the same as reading your texts or listening to your calls. NYPD officials told the NY Daily News that FalcoNet is not used to intercept calls or collect information such as text messages from phones not involved in an investigation. The system is portable: it can ride on a police car, mount on a drone or travel in a backpack. Cognyte specific AI capabilities within FalcoNet remain incompletely documented in public materials, in part because the NYPD’s February 2026 policy quietly deleted an earlier statement that cell-site simulators do not use artificial intelligence or machine learning.
The safeguards on paper
The NYPD’s published policy requires a warrant in most cases, but carved out exceptions and limited access rules define the operational boundaries.
Under the department’s published impact-and-use policy, investigators must obtain a judge-issued search warrant backed by probable cause before deploying a cell-site simulator in most cases. Exigent circumstances involving an immediate threat of serious injury or death permit the use of prior warrants, with a retroactive warrant generally required within 48 hours. No warrant is needed at all to locate a missing or suicidal person.
The operation is limited to four TARU members who run the system through laptops on a closed, stand-alone network. NYPD officials pointed to the February rescue of a 15-year-old girl kidnapped from a Queens group home as an example of what quick phone location skills can accomplish. The reference is to phone tracking technology in general, and no source confirms that FalcoNet itself was involved in that case.
The written framework sounds reasonable on paper. The gap between what a policy promises and what anyone outside the department can independently confirm is where the real problem begins.
Before the NYPD had a written cell-site simulator policy or a warrant requirement, it outlasted the technology 1,000 timesaccording to a report by the NY Daily News. This history is precisely why advocacy groups treat current insurance policies as a starting point rather than a conclusion.
The New York Civil Liberties Union raised concerns that the cell-site simulators could capture identifying data from phones belonging to people who are not investigative targets. This concern joins another: The Digital Forensics Unit of the Legal Aid Society noted that written restrictions on content collection cannot be independently verified without audit records or external review. The Surveillance Technology Oversight Project pointed out that the NYPD does not report the use of cell-site simulators on a case-by-case basis, even when it publishes reports on drone deployments. This asymmetry makes it difficult to judge who the department is targeting or whether its own policies are being followed.
Cognyte did not respond to requests for comment, according to the NY Daily News. Human rights organizations have previously raised concerns about Cognyte products being sold to governments, accusing them of targeting dissidents and political opponents. These charges are attributed, not judged.
Before this contract is approved, the NYPD should commit to aggregate or case-level usage reports, clear data protection and erasure rules, and a credible path to independent audits. Warrant requirements and TARU-only access are a floor, not a ceiling. The standard for deploying AI-assisted surveillance over one of the world’s densest cities should be transparency that the public can actually measure.