Gemma speaks to ANFAIA community
July 13th, 2026— ANFAIA
Eticas.ai's Founder & CEO, Dr. Gemma Galdón-Clavell, recently sat down with a group of young students from the ANFAIA community for a conversation about her professional journey and the case for auditing artificial intelligence. The full session is in Spanish — for readers who don't speak the language, here's an English recap of the main themes she covered.
The "sociotechnical layer" AI keeps missing
Gemma's starting point is that AI is too often built around what she calls "pure anatomy": using massive historical datasets to find patterns and produce an output. What that approach leaves out is the sociotechnical layer: the real-world social dynamics that shape how a system actually behaves once it's deployed. In technical terms, minority patterns get treated as errors or outliers. That might be a reasonable simplification for a GPS. In a high-risk domain like medicine, dismissing a minority symptom as noise can be dangerous. Her point isn't only an ethical one, either: building in that social complexity, she argues, makes AI systems perform better and gives organizations a real competitive edge, not just a cleaner conscience.
Why AI has to be audited where it actually runs
Gemma draws a comparison to aeronautical engineering: an engineer can design a plane that looks perfect on paper without ever checking whether it takes off or stays in the air. Most AI evaluation today works the same way — tested in the lab, rarely checked in production. Technology behaves unpredictably once it meets real people in real conditions, and auditing systems as they actually run is how those risks get caught before they cause harm.
A finding that hits close to home for education
Among the practical examples she shared, one lands squarely in the education space: her research suggests that AI tools built for children should sound less empathetic and more "robotic." The reasoning is straightforward — a more clinical, tool-like tone helps prevent kids from misusing the system or treating it like a friend rather than the technical tool it actually is.
Rescuing AI from itself
Gemma closes by framing her work not as a rejection of AI, but as a corrective to the arrogance of a purely technical approach that ignores social consequences. She describes Eticas.ai as her legacy — a mission-driven firm built to make technology accountable and safe for the people it affects.
Source URL (in Spanish): www.youtube.com/watch?v=hPVXCjklhso