Skip to content
Submit incident
AII-1002-13651
ReportedReported by fewer than three independent publishers, with no official finding.Supporting sources are pending verification.

Spanish Man Loses Retirement Pension After Following ChatGPT Advice

A Spanish man seeking early retirement consulted ChatGPT to calculate his pension and guide his application, and the system hallucinated a non-existent law provision to justify an 800-euro monthly estimate. His actual approved pension was 200 euros per month, resulting in a permanent loss of 600 euros monthly.

3 September 2025 · OpenAI · Reliability and hallucination
3 September 2025
AII-1002-13651
Quick facts
Organization
OpenAI (subject)
Category
Reliability and hallucination
Country
Spain
Date
3 September 2025
Drafted by
Raidu Incident Detection

What happened

A 60-year-old Spanish man enrolled in a Special Regime sought to retire early and consulted ChatGPT to calculate his expected pension and steps to submit his application to Social Security.

ChatGPT cited a non-existent "disposición cuadragésima séptima" (forty-seventh provision) to justify a calculation predicting a monthly pension of approximately 800 euros.

The man followed ChatGPT's guidance without consulting a professional or Social Security directly, and submitted his retirement application.

Social Security approved a monthly pension of only 200 euros—one-quarter of the amount ChatGPT predicted.

Reported impact

The man's actual retirement income was reduced to 25% of what ChatGPT indicated he would receive, resulting in a permanent loss of 600 euros per month throughout his lifetime.

Labor attorney Ignacio de la Calzada, who shared the case, stated the error had cost his client "una pasta de por vida" (a fortune for life).

Index analysis

De la Calzada noted that ChatGPT invented false legal provisions and that the tool is unreliable for consequential legal decisions: "Se inventó una barbaridad" (it made up something absurd).

The attorney emphasized that ChatGPT lacks the specificity and reliability needed for professional advice and repeatedly warned against trusting the system with decisions affecting the future.

Governance controls

De la Calzada recommended consulting qualified professionals for legal and administrative matters rather than relying on AI systems, stating "Usad siempre un profesional" (always use a professional).

Classification

Organization
OpenAI (subject)
Category
Reliability and hallucination
Industry
Financial and insurance services
Country
Spain
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Economic or property
Harmed
Consumers
Failure modes
Fabricated citation, Hallucination, Overreliance, Unsafe recommendation
AI systems
ChatGPT

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident