{"resourceId":"ncaa-signify-threat-matrix-2025-26-report","versions":[{"version":"external-e8dff43599f9f2c8ab3889b3fab853221f505dad900ae243f4dc64219ffccdc9","resource":{"id":"ncaa-signify-threat-matrix-2025-26-report","title":"NCAA 2025/26 Threat Matrix report: AI flagged 61,724 messages, analysts verified 7,666 as abusive, and 4 cases went toward law enforcement","organization":"Signify Group for the NCAA","sector":"Student-athlete online safety and sports-betting harassment monitoring","geography":"United States (national NCAA championships)","publishedAt":"Posted October 1, 2026 (file date)","publicationDate":"2026-10-01","eventDate":null,"sourceName":"NCAA (report by Signify Group)","sourceLabel":"Vendor-produced monitoring report published by the NCAA","sourceUrl":"https://ncaaorg.s3.amazonaws.com/ncaa/wagering/2025-26NCAA_ThreatMatrixPublicReport.pdf","evidenceClass":"vendor-claim","outcomeClass":"mixed","topics":["data-security","operating-model","governance-procurement"],"finding":"Across nine sports in 2025/26, Threat Matrix scanned 977,457 messages mentioning monitored accounts; AI flagged 61,724 for analyst review, and 7,666 were verified abusive and reported to platforms, about one in eight flags. Basketball accounted for 6,758 verified posts. Signify investigated 127 accounts, identified 18 users and recommended 4 cases for law enforcement referral. Abuse attributable to gamblers rose to 7.9% from 2.2% the prior year, and the report says AI image editing adds to sexualized abuse of female athletes.","sledRelevance":"Why now: the 2025/26 report was posted Oct 1, 2026 and succeeds the archived 2024/25 report with a third year of data, expanded coverage and new betting-attribution analysis. It is one of the few athletics AI deployments that publishes its funnel from AI flags to verified cases.","evidence":"Annual report by the monitoring vendor, published on the NCAA site. Counts by sport for messages, AI flags reviewed by analysts, and verified abuse; investigation and referral counts; year-on-year comparisons with category changes noted. Publication date inferred from the file's server and document metadata (Oct 1, 2026). No independent audit, no false-negative estimate, and recall cannot be assessed.","architectureImplications":"AI detection over public social posts in 56 languages, followed by analyst review, OSINT investigation and case packs. Interpretation: the 61,724-to-7,666 funnel shows the human review load any athletics AI flagging system creates.","governanceImplications":"Case packs and law-enforcement referrals involve identifying individuals; NCAA and schools need thresholds, evidence standards and athlete-welfare follow-up for each step.","securityPrivacyImplications":"Monitoring and unmasking accounts uses open-source intelligence on members of the public; retention and sharing of identity data with schools and law enforcement need clear rules.","caveats":"Vendor-produced; measures only public posts on monitored accounts during championships; category definitions changed year to year; no false-negative or accuracy audit; publication date inferred from file metadata.","streamIds":["college-athletics"],"roles":{"sales":"Sales framing: this report shows what an athletics AI monitoring service actually produces: a large review queue, a smaller set of verified cases and a handful of referrals. A bounded offer is an athlete online-safety response review for a department or conference, covering who acts on case packs and how athletes are supported. Do not claim AI stops abuse or catches all threats; the report has no recall measure.","engineering":"Engineering framing: the useful numbers are the funnel. AI flagged about 6% of scanned messages, and analysts confirmed about 12% of flags. Engineers evaluating similar tools should ask for the same breakdown, plus language coverage, image handling and how flags reach staff. Proof means a pilot reporting flags, confirmed cases and analyst hours per event, with a sample audit of unflagged messages.","delivery":"Delivery framing: monitoring creates work for campus staff once cases are confirmed. Delivery should define who at the school receives an NCAA case pack, how athlete welfare staff contact the athlete, when campus police or event security are involved, and how outcomes are recorded. The report itself recommends a fast-track intelligence process with event security. Interpretation: schools should test that handoff before championship season."},"roleDetails":{"sales":{"customerProblem":"Athletes face online abuse and betting-related threats, and departments lack a clear process to act on monitoring results.","stakeholders":["Senior woman administrator and athlete welfare staff","Compliance director","Campus police and event security","Conference office","Mental health providers"],"discoveryQuestions":["Who receives abuse or threat alerts about your athletes today?","How are affected athletes contacted and supported?","When are campus police involved?","Do you monitor regular-season events?"],"valueHypothesis":"A defined response process may turn monitoring alerts into timely athlete support and appropriate referrals.","potentialEngagement":"An online-safety alert response review and tabletop exercise for one department or conference.","unsupportedClaims":["Detection of all abuse","Reduction in abuse caused by monitoring","Independent accuracy audit","Regular-season coverage"]},"engineering":{"fit":"Conferences and departments considering social-media abuse monitoring for athletes.","architectureIntegration":"Vendor AI scanning of public social posts, analyst review, and case packs delivered to the NCAA.","prerequisites":["List of monitored accounts with consent","Alert routing to staff","Case tracking"],"constraints":["Public posts only","Championship-period scope","Platform policy changes affecting visibility"],"security":"Protect case packs containing identity information; restrict and log access; set retention.","proofOfValue":["Flags, confirmed cases and analyst hours reported per event","Sample audit of unflagged messages","Time from confirmed case to athlete contact"]},"delivery":{"work":"Connect monitoring outputs to campus athlete welfare and security processes.","dependencies":["NCAA or conference monitoring program","Campus police","Welfare and mental health staff"],"ownership":"Athlete welfare lead with compliance and campus police.","skillsAdoption":"Train staff to read case packs and to support athletes after abuse.","governanceCheckpoints":["Referral thresholds agreed","Athlete consent for account monitoring","Season review of cases and outcomes"],"acceptanceCriteria":["Proposed: every confirmed case involving a school's athlete reaches a named staff member within a set time","Proposed: outcomes recorded for all referred cases"],"risks":["Case packs not acted on","Privacy harm from unmasking errors","Athletes unsupported after abuse"]}},"retrievedAt":"2026-10-07T22:12:36-05:00","enrichedAt":"2026-10-07T22:16:06-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Confirmed cases create work for welfare, compliance and security staff; analysts carry the review load on the vendor side.","procurementImplications":"Ask monitoring vendors for flag-to-verified funnels, language coverage, recall audits and data-sharing terms.","operatingModelImplications":"AI monitoring feeds human investigation and campus response, so campus staffing determines the real-world effect.","sourceVerification":{"responseId":"curl-ncaa-threatmatrix-2025-26-20261007T2213","openedUrl":"https://ncaaorg.s3.amazonaws.com/ncaa/wagering/2025-26NCAA_ThreatMatrixPublicReport.pdf","referenceExcerpt":"Across all monitored events in the 2025/26 season, 127 unique accounts shared content that met investigation thresholds","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}