VioGén risk-assessment used by Spanish National Police misclassified victims

An academic review and investigative reporting documented transparency, accuracy, and governance problems with VioGén, the Spanish police risk-assessment tool overseen by the Interior Ministry. Reporting and analyses found that the system classified many cases as negligible or low risk and that some victims later suffered repeat attacks or were killed, prompting rights and oversight concerns.

Spanish National Police (Ministry of the Interior) · Incident Nov 19, 2022 · Indexed Jun 10, 2026 · 2 sources

Incomplete input data and an opaque, non-auditable scoring system caused VioGén to underestimate risk for some victims.
What
An academic review and investigative reporting documented transparency, accuracy, and governance problems with VioGén, the Spanish police risk-assessment tool overseen by the Interior Ministry.
Incident date
Nov 19, 2022
Who
Spanish National Police (Ministry of the Interior)
Failure mode
Hallucination
AI surface
Algorithmic Decision
Severity
High

What happened

An academic review and major investigative reporting documented problems in VioGén, the Spanish National Police system used to score the risk of repeat intimate-partner violence. Investigations found that VioGén labeled many victims as negligible or low risk while a subset of those later reported repeat assaults or were killed after assessment. The academic review flagged shortcomings in transparency, accuracy testing, data access, and governance that could increase rights risks for people evaluated by the system.

What broke inside the model

Failure path · mode profile · Hallucination
  1. 01 · TriggerA user asks for a fact, a citation, or a figure.
  2. 02 · Model stepThe model writes a fluent, confident answer.
  3. 03 · Control gapNothing ties the claim back to a real source.
  4. 04 · FailureA fabricated fact ships as if it were verified.
  5. 05 · ConsequenceThe false claim reaches a customer, a court, or the public.

Confidence holds, and even spikes, as the claim detaches from any source.

Failures were driven by a combination of incomplete or low-quality input data from hurried or partial police interviews, limitations of the standardized questionnaire used to generate scores, and the system’s opaque design and limited external auditability. That combination produced systematic underestimation of risk in some cases and reduced effective human oversight of protective decisions.

Public visibilityHigh
Regulatory exposurePossible
Customer impactClass-wide
Financial impactUnknown
Time to disclosureMonths
  1. PressAn Algorithm Told Police She Was Safe. Then Her Husband Killed Her.nytimes.com
  2. PrimaryThree predictive policing approaches in Spain: Viogén, RisCanvi ... (academic review, PDF)regulation.blogs.uv.es
Permalinkhttps://failureindex.ai/failures/viog-risk-assessment-used-spanish-national
CitationAI Failure Index. "VioGén risk-assessment used by Spanish National Police misclassified victims" (FI-0455). Realm Labs. https://failureindex.ai/failures/viog-risk-assessment-used-spanish-national (indexed Jun 10, 2026).
Share cardA branded image of this record for posts and slides.

Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0455. Full dataset at /data.

Note from Realm Labs, the Index steward

How Realm fits

Controls for this failure mode
  • Prism
  • OmniGuard
  • AI Detection & Response (AIDR)

This entry sits in the index's predictive wing: a system that scores, ranks, perceives, or steers rather than generates. Realm's runtime layer is built for the generative and agentic systems now moving into these same decision seats, where it watches a model's internal state and holds an unsupported claim or an unchecked action before it commits. The control gap on this record, an automated decision that reached people with no runtime check in front of it, is the same gap. The index keeps predictive failures on the record because the pattern carries straight into the systems shipping today.