We use AI every day, but we don't trust the people who build it: the paradox according to Enrique Dans

Note: the content downloaded for this article is a paywall teaser (Medium flags it as a "member-only story") and the text cuts off literally mid-sentence, with an incomplete figure ("71% …"). This summary is therefore necessarily brief and limited to what the excerpt actually says, without completing or inventing the missing data.

The article starts from an MIT Technology Review piece titled "People really hate AI, so why can't they get enough?", which poses an apparent paradox: despite the distrust, fear and anger AI generates among broad swaths of the population, its use keeps growing. The example cited is that of students who have booed AI-related speakers and protested at graduation ceremonies over its impact on the job market while, at the same time, using ChatGPT, Gemini or Claude ever more frequently.

The author's reading is that this is neither hypocrisy nor a new form of tech addiction, but something simpler: it is possible to value a tool and at the same time deeply distrust those who build it, how they intend to insert it into people's lives and the consequences of that process. In other words, accepting the product's usefulness does not imply approval of the industry or its intentions.

The available excerpt offers several figures to illustrate that gap between use and trust: according to the Pew Research Center, 49% of US adults already use some chatbot and 24% do so daily, while 63% believe AI is advancing too fast. The OECD estimates that more than a third of citizens in its 38 member countries used generative AI during 2025. Stanford's AI Index report, for its part, notes that 59% of those surveyed believe the benefits outweigh the drawbacks, although 52% say they are nervous about the technology. The text also begins to cite a University of Melbourne and KPMG study focused on Spain, according to which 72% of Spaniards accept or approve of AI, but the download breaks off just as a second percentage (71%) is introduced, with no way to know exactly what it refers to, so that figure cannot be reproduced reliably.

Taken together, the available fragment suggests that the article's central argument is to distinguish between the practical adoption of AI, which is on a clear upward curve in several regions of the world, and trust in the industry producing it, which remains low or ambivalent; but since this is a paywalled preview, the full development of that idea, as well as the author's final conclusions, is not available in the content received.