Real Madrid is facing intense scrutiny after a former staff member alleged the club relied on the free version of ChatGPT to determine player supplement regimes. Itziar Gonzalez de Arriba, who served as a physiotherapist and nutritionist from 2021 to 2025, took to social media this March to expose the practice. Her comments follow a separate scandal where medical personnel treated Kylian Mbappé's healthy knee instead of the injured one, leading to the dismissal of the entire medical team in January.

For data engineers, this incident highlights a catastrophic failure in AI governance. Gonzalez noted she was unsure what was worse: the surgical error or the reliance on a public large language model for medical decisions. Using general-purpose AI for specialized health protocols ignores basic safety guardrails. Free tiers lack the fine-tuning, privacy controls, and verification necessary for clinical environments. They are designed for conversation, not prescription.

Gonzalez criticized the decision-makers as connected but unqualified, hoping replacements prioritize competence over convenience. Mbappé avoided surgery only after a French physician intervened. This situation serves as a stark reminder for ML teams deploying models in high-stakes sectors. When organizations bypass specialized tools for consumer-grade AI, the risk shifts from theoretical hallucination to physical harm. As AI integration accelerates across industries, Real Madrid's controversy underscores the necessity of rigorous validation pipelines before putting model outputs into production workflows. Responsible deployment means knowing exactly where not to use generative models. The boundary between assistance and autonomy must remain clear, especially when human health is on the line.

Source: RIA Novosti