Let us know when you are using this technology
En España, la inteligencia artificial ha dejado de ser una abstracción técnica para convertirse en una pregunta moral y política de primer orden. La primera encuesta del CIS dedicada íntegramente a la IA revela que una sociedad que ya convive con la tecnología —casi la mitad ha usado ChatGPT— exige, con abrumadora mayoría, saber cuándo las máquinas sustituyen a las personas y quién vigila ese proceso. Es la voz de una ciudadanía que no rechaza el futuro, sino que negocia las condiciones en que está dispuesta a habitarlo.
- El 93% de los españoles exige que las empresas declaren cuándo usan IA en lugar de personas, una demanda que convierte la transparencia en imperativo social antes de que sea obligación legal.
- La incertidumbre domina el estado de ánimo: tres cuartas partes de quienes conocen la IA confiesan sentirse inseguros, y más de la mitad teme que destruya más empleo del que crea.
- Los riesgos no son abstractos: los deepfakes no consensuados difundidos en Almendralejo pusieron rostro concreto al miedo, y el 86,9% ya asocia la IA con la propagación de desinformación.
- La privacidad se perfila como línea roja: el 76,8% rechaza que las empresas vendan datos personales, y tres cuartas partes denuncian que las políticas de privacidad son demasiado opacas.
- Frente al temor, emerge una confianza selectiva: medicina, industria y agricultura concentran las esperanzas de una ciudadanía que no condena la tecnología, sino que exige gobernarla.
El Centro de Investigaciones Sociológicas publicó este mes su primera encuesta dedicada en exclusiva a la inteligencia artificial, y el retrato que ofrece es el de una sociedad en tensión entre la adopción y la desconfianza. El 41,1% de los españoles ha usado ChatGPT al menos una vez en el último año, y el 92,3% conoce las aplicaciones de IA generativa. Sin embargo, esa familiaridad no ha traído tranquilidad: tres cuartas partes de quienes conocen la tecnología reconocen que la incertidumbre es su sentimiento predominante, y solo el 27,4% la contempla con optimismo.
La encuesta, realizada entre el 6 y el 15 de febrero con 4.004 participantes, recoge miedos precisos y arraigados. El 86,9% teme que la IA facilite la difusión de desinformación; el 80,6%, que sirva para cometer delitos. Estos temores no son puramente hipotéticos: la circulación reciente de vídeos deepfake de contenido sexual no consentido en Almendralejo dejó una huella visible en la conciencia pública. A ello se suma la preocupación por el empleo —el 56% cree que la IA perjudicará más que beneficiará al mercado laboral— y por la privacidad, con un 76,8% que rechaza que las empresas comercialicen datos personales.
Pero la encuesta también dibuja una ciudadanía pragmática, no apocalíptica. El 66% ve en la medicina el campo donde la IA puede aportar más beneficios; la industria y la agricultura le siguen de cerca. Y las demandas mayoritarias no son de prohibición, sino de control: el 92,7% exige que las organizaciones informen cuando usen IA en lugar de trabajadores humanos, y el 93,4% reclama que el desarrollo de la tecnología esté regulado.
Lo que emerge de estos datos es una negociación colectiva en curso. Los españoles no han dado la espalda a la inteligencia artificial; han fijado las condiciones bajo las cuales están dispuestos a convivir con ella: transparencia sobre su uso, protección de los datos que la alimentan y un marco de gobernanza que garantice que alguien, efectivamente, está al mando.
Spain's Center for Sociological Research released its first survey devoted entirely to artificial intelligence this month, and the results paint a portrait of a population caught between curiosity and unease. Nearly half of all Spaniards—41.1 percent—have used ChatGPT at least once in the past year. Yet the same country that has embraced this technology overwhelmingly demands to know when companies are using it. Ninety-two point seven percent of those surveyed believe organizations must disclose when they deploy AI instead of human workers. An even larger majority, 93.4 percent, insists that AI development itself requires regulation.
The survey, conducted between February 6 and 15 among 4,004 respondents, captures a moment when artificial intelligence has moved from the margins of public conversation into its center. Most Spaniards—92.3 percent—have at least heard of generative AI applications like ChatGPT. But awareness has not bred comfort. Three-quarters of those familiar with AI report that uncertainty is their dominant feeling. More than half say they are afraid. Only 27.4 percent view the technology with optimism. When asked about its impact on employment, 56 percent believe AI will cause more harm than good to the labor market.
The sources of this anxiety are concrete and specific. Spaniards recognize that AI can be weaponized. Eighty-six point nine percent worry it will spread disinformation and false claims. Eighty point six percent fear it could facilitate crime or illegal acts. The survey was conducted against a backdrop of real incidents—deepfake videos of non-consensual sexual imagery had recently circulated in Almendralejo, a town in western Spain—and the public's concern reflects that lived awareness of the technology's darker uses.
Privacy emerges as another central worry. Seventy-six point eight percent strongly agree that companies should be prohibited from collecting personal data for sale to third parties. Three-quarters say privacy policies are too opaque. The message is clear: Spaniards want both transparency about when AI is being used and protection of the data that trains it.
Yet the survey also reveals where Spaniards see genuine promise. Medicine tops the list—66 percent believe AI could deliver significant benefits there. Manufacturing follows at 62.3 percent, and agriculture at 51.8 percent. These are sectors where the technology's capacity to process information and identify patterns aligns with concrete human needs.
AI itself is not new to Spanish daily life. Navigation apps, email filters, and streaming recommendations all rely on it. But the arrival of ChatGPT in November 2022 marked a turning point. For the first time, a single tool demonstrated what generative AI could do—write coherent text, sustain conversation, create content from simple instructions. The technology moved from invisible infrastructure to visible presence. In the two years since, countless applications have emerged capable of generating increasingly realistic text, images, video, and audio.
This visibility has sparked both wonder and alarm. Some experts and executives have called for pausing AI development altogether; that 2023 manifesto reached only 24 percent of Spaniards. But the public proved far more aware of Hollywood's 2023 writers' strike and a 2024 letter from major musicians defending their copyright—events framed around protecting human creators and workers. The survey suggests Spaniards are thinking less about existential risk and more about immediate, tangible harms: their jobs, their privacy, their ability to trust what they see and read.
What emerges is a pragmatic demand: let us know when you are using this technology, protect our data, and make sure someone is actually governing how it develops. The Spanish public has not rejected AI. It has simply insisted on conditions for living alongside it.
Notable Quotes
Three-quarters of those familiar with AI report that uncertainty is their dominant feeling, with only 27.4 percent viewing the technology with optimism.— CIS survey findings
The Hearth Conversation Another angle on the story
Why does it matter that 93 percent want disclosure? That's a huge number—almost consensus.
Because right now, companies don't have to tell you. You might be talking to an AI without knowing it. You might be rejected for a loan based on an algorithm you never saw. The demand for disclosure is really a demand for visibility—for the right to know when a machine is making decisions about you.
But 41 percent have used ChatGPT. Doesn't that suggest people are comfortable with AI?
There's a difference between using a tool and trusting a system. People will experiment with ChatGPT for fun or convenience. That doesn't mean they want it making hiring decisions or filtering their news. The comfort is shallow.
The survey mentions deepfakes in Almendralejo. How much is fear based on real incidents versus abstract worry?
Both. The deepfakes were real—non-consensual sexual imagery created by AI. That's not theoretical harm. But the broader anxiety is also real. People understand that if a tool can create convincing fake video, it can be used to spread lies at scale. They've seen it happen elsewhere.
Why do you think medicine and agriculture scored highest for potential benefits?
Because those are domains where AI's strength—processing huge amounts of data, finding patterns humans miss—directly solves a problem. A doctor needs help reading scans. A farmer needs to know when to irrigate. The technology serves a clear human need. Labor displacement is less obvious there, or at least less immediate.
The survey says only 24 percent heard about the 2023 manifesto calling for a pause on AI development. Why did that message not land?
Because it felt abstract. A pause on development? For what? How? By whom? Meanwhile, the writers' strike was concrete—people lost jobs, shows didn't get made. That's a story people understand. The public is thinking about immediate, visible consequences, not hypothetical futures.