Our Take
Thomson Reuters is claiming it sells public records, not surveillance tools—a distinction that collapses under scrutiny of how CLEAR actually works in ICE's hands.
Why it matters
Thomson Reuters data powers immigration enforcement at scale. The shareholder push for transparency matters now because ICE enforcement is intensifying and the company's employees and customers deserve clarity on what the platform enables.
Do this week
Legal tech vendors and procurement teams: audit contracts with data brokers and enforcement agencies—specifically what inferences their platforms enable—before renewal or expansion.
Shareholders Vote on Thomson Reuters Transparency
The BC General Employees' Union, a minority shareholder in Thomson Reuters, has filed a resolution asking the company to commission and publish an independent human rights impact assessment of how its products are used by law enforcement and immigration agencies. The vote occurs next week. The push builds on shareholder engagement that began in 2020, though Thomson Reuters' relationship with U.S. Immigration and Customs Enforcement predates that.
Thomson Reuters signed contracts with ICE starting in 2017 when the agency launched its "Extreme Vetting" program. The deals are valued at tens of millions of dollars. ICE has described the company's data services as "mission critical" to its operations and has integrated them into tools including Motorola's license plate readers, PenLink's location tracking, and Palantir's software.
The company's public position is narrow: it provides access to records that are already public, not records requiring a warrant, and therefore does not supply surveillance tools. Two employees who raised concerns internally were allegedly terminated, prompting a whistleblower lawsuit and an unfair labor practice complaint. Hundreds of current Thomson Reuters employees signed a letter calling for greater transparency.
What CLEAR Actually Does in Immigration Enforcement
Thomson Reuters' argument conflates individual data points with aggregated surveillance. CLEAR retrieves thousands of data points about any individual: current and historical addresses, known associates, vehicle information, employment history, phone numbers, financial records, and location history. The platform maps social networks and aggregates data from social media and, according to published reporting, dating profiles. Its power lies not in any single field but in the instant assembly of these traces into a profile.
Thomson Reuters notes that CLEAR does not include an immigration status field. That is technically accurate but incomplete. ICE agents do not need a labeled field to make immigration determinations. CLEAR exposes inferences: use of an Individual Taxpayer Identification Number instead of a Social Security Number, residence in states requiring citizenship proof, regular international wire transfers, absence from voter rolls, employment in cash-wage sectors, and prepaid phones instead of credit products. The platform also provides immigration court proceeding records, visa records, I-94 arrival and departure records, and E-Verify non-confirmation records. It generates risk scores, confidence scores, and verification failure flags for individuals who cannot be fully verified through standard U.S. documentation, and its network analysis tools map connections between flagged and unflagged individuals.
An ICE agent entering a name or address can generate a profile showing no SSN-associated credit history, ITIN use, residence at known worker housing, employment in agriculture, and regular contact with someone with immigration court records. The absence of a labeled immigration status field does not diminish utility—it requires one inferential step.
Why the Distinction Matters for Data Governance
The Thomson Reuters case exposes a structural gap in how corporations frame data product governance. The company treats "repackaging public records" as a neutral activity distinct from "providing surveillance tools." That framing breaks down when the product's purpose is to help users "locate and identify individuals" by fusing thousands of data points into a profile revealing far more than any single point would suggest.
Organizations procuring data services or selling access to government agencies should examine their own contracts and platform capabilities against this model. The question is not whether individual data points are public—it is whether the aggregation, inference, and inference speed create a new capability that shifts the nature of the activity. A human rights impact assessment would clarify that distinction or expose that it does not exist.