SanctionsKit’s study of 96 selected company searches found 17 incorrect AI search summaries, including three that described listed companies as absent from an official U.S. sanctions list.
Denver, CO, Oct. 20, 2026 (GLOBE NEWSWIRE) — SanctionsKit LLC has announced findings from its sanctions screening study after identifying 17 incorrect AI-generated search summaries across 96 selected company searches. Three answers said a company was absent from the official list when records showed it was listed. The study tested one search engine against one official U.S. sanctions list on October 8, 2026.
An AI search summary said “no public listing,” but the official sanctions list showed a match. SanctionsKit found 17 incorrect answers across 96 selected company searches on October 8, 2026.
A wrong answer can send a business decision in the wrong direction. An incorrect “not listed” answer could cause an employee to skip further review of a listed company. An incorrect “listed” answer could delay customer onboarding or disrupt a supplier relationship unnecessarily. An uncertain answer could also be mistaken for “no match.” These risks matter when employees rely on quick web searches without a defined sanctions screening process.
Alex Smith, the research author at SanctionsKit, said: “Employees need a clear process that checks official records, sends possible matches for review, and keeps the evidence behind each decision.”
In one example, an AI search summary said there was “no public listing” for a company whose exact name and country appeared in the official file checked that day. The answer attributed its conclusion to official records, but those records showed the company was listed.
Another answer said a company was still listed while citing the official notice that recorded its removal. Researchers followed that citation and checked the current list. The company was absent. The answer carried a government citation, but its conclusion was wrong.
Of the 17 incorrect answers, 14 called removed companies still listed and three called listed companies absent. The study did not track employee decisions or resulting business harm.
SanctionsKit helps businesses establish a clear screening process with sanctions screening software built on official-source data and defined matching rules, without AI. Employees can screen customers and suppliers through its web interface. Businesses can also use its sanctions screening API to bring automated sanctions screening into customer onboarding, supplier due diligence and internal applications. Training employees to use these tools gives them a clear procedure to follow when a screening decision matters.
Each saved check records the source versions and matching policy used, creating a reproducible record of what was checked. Matching profiles help teams balance sensitivity and false flags. Saved potential matches enter the Review queue, where employees can inspect official-source evidence, assign follow-up work and document their decisions. This gives businesses a clear distinction between a potential match awaiting review and a completed review, with an audit trail connecting the original check to the recorded decision.
The platform monitors source refresh schedules and alerts its team to late or failed updates so they can respond promptly. If a required source is unavailable, the API returns an error. An incomplete check does not become a successful no-match result.
The full AI sanctions screening study identifies the search engine, source list and companies tested. It includes the captured answers, official records, methods and downloadable charts. Separate API experiments examine how models respond under different conditions. Journalists can inspect the evidence behind the findings and contact SanctionsKit about its methods.
The test covered 64 removed companies and 32 listed companies, with one search per company. Besides the 17 incorrect answers, 68 gave the correct status, 10 did not establish current list status, and one search had no AI summary. Uncertain answers were not counted as errors. This selected sample does not establish an overall error rate or determine whether a transaction is lawful.
About SanctionsKit LLC
SanctionsKit provides sanctions screening software and an API for businesses. Its platform combines customer and supplier screening, review workflows, ongoing monitoring, and audit records. Teams can inspect official-source evidence, investigate potential matches, and document decisions, with source versions and matching rules recorded for each saved check. Learn more at www.SanctionsKit.com.
Press Inquiries
Alex Smith
support@sanctionskit.com
https://www.sanctionskit.com
