> ## Content Index > Fetch the complete content index at: https://www.thecriticalloop.com/llms.txt > Use this file to discover other available public pages before exploring further. # FROM COCKPITS TO ALGORITHMS: THE AVERAGE HUMAN DOESN'T EXIST. SO WHY ARE WE BUILDING AI FOR ONE? - URL: https://www.thecriticalloop.com/the-average-pilot-is-back/ - Published: 2026-10-28T15:24:10.000Z - Updated: 2026-10-29T21:17:57.000Z - Description: AI is the first technology that could fit itself to every single person. So why are we using it to rebuild the average? Ricardo Wagner examines the under-utilised power of technology and highlights the 3 things leaders should take back to the office on Monday. - Author: Ricardo Wagner - Tags: Public-life, Ricardo Wagner # By Ricardo Wagner, 29 October 2026 In the late 1940s, the United States Air Force had a problem nobody could explain. It had just entered the jet age, and its aircraft were falling out of the sky. The planes were new, the pilots were among the best in the country, and the investigations kept finding neither mechanical failure nor pilot error. On the worst day, seventeen aircraft went down, and none of them was in combat. Eventually someone thought to look at the cockpit itself. It was designed in 1926 around the average body measurements of pilots of that era. In 1950, the Air Force measured 4,063 pilots, and a 23-year-old lieutenant called Gilbert Daniels took ten of those body dimensions and asked a deceptively simple question: how many of them actually fitted the average? The answer was none. Not one. Even using just three of those dimensions, fewer than 3.5 per cent fitted. The average pilot was a statistical fiction, and a cockpit built for him fitted nobody. The fix was not better pilots. It was to make the cockpit adjust to the person: adjustable seats, pedals and helmet straps. The accidents fell sharply. It is worth remembering the next time you slide your car seat forward without a second thought. Seventy-five years later, the average pilot is back. This time he lives inside our organisations, our data and, increasingly, our AI systems. ## **AI as a mirror** It is tempting to blame AI for a new set of problems. But AI has not invented the dilemmas of our time; it has reflected them at us, at scale. Look into that mirror and three familiar faces will appear: **1\. The leader seeking answers** Most of us were educated in a model built for the industrial age. Desks in rows, a bell, and an exam with a single correct answer. We were trained to have answers, not to ask questions. Organisations then did the rest. They promoted the person who always knew, and that person eventually reached the boardroom. The turnstile became a KPI; the foreman became a dashboard. We swapped iron for silicon and kept the factory mindset intact. That mindset now meets a technology that rewards the opposite behaviour. Working well with AI means experimenting, getting it wrong, testing again, and doing so in front of others. The data from organisations shows who struggles most. In the finance function of one multinational healthcare company this year, 95 per cent of staff used AI tools every month. When usage was broken down by seniority, the lowest group was the directors, at roughly 40 per cent of the median use of their own teams. In the same team, three analysts did not wait for certainty; they tried things. One reduced a monthly process from more than 90 minutes to 30 minutes. After giving time back to 8 colleagues, they then taught everyone else. When the directors finally started using the tools, the team followed. Maturity is contagious and has no dependency on job title. **2\. The one who fits the norm:** The people who pay most for designing around the average are those who never fit it, and above all, people with disabilities. The [World Health Organisation (WHO) ](https://www.who.int/news-room/fact-sheets/detail/disability-and-health?ref=thecriticalloop.com)estimates that more than 1.3 billion people live with a significant disability. That’s about 16 per cent of the world. It is hard to see how anyone could still call that a niche. They are your customers, your colleagues and your shareholders. Accessibility shows that the problem is not the person in the wheelchair but the staircase. People are not disabled in the abstract; environments and tools disable them and determine who fits, who belongs and who gets to take part are design decisions and, therefore, choices. For the first time in history, we have a technology that can adapt to each individual. It can take into account their pace, their language, their way of reading the world. We no longer need to depend on averages. Yet most AI deployments are still tuned to a default user, measured against a typical case and evaluated on the median. We have been handed an adjustable cockpit and bolted the seat back into place. But those who notice this irony gain an advantage. Let’s take, for example, the automatic captions. They were originally built for people who cannot hear. But they also allow executives to follow a meeting in a departure lounge without headphones. It was not technology that taught people to communicate. It was people outside the norm who taught technology to listen. Organisations that start from the centre end up having to make adjustments later. The ones that start from the edge learn sooner. 1. **The one traded for gold** In the colonial history of the Americas, mirrors and glass beads were exchanged for gold. We tend to remember that as other people's naivety. Today we are the gold, and the mirror is the screen: the free app, the convenient login, the flattering filter. Very few of us have read the last terms of service we accepted in full. The shift is subtle but profound. Industrial society turned people into averages to organise them. Algorithmic society turns people into inferences to anticipate them. The well-known case of a large American retailer that sent baby-product coupons to a teenager before her own father knew she was pregnant involved no stolen password and no confession. It involved inference. In places like Brazil, where I work, when you go to the cashier at every chemist, they ask if you would like your tax number on the receipt. Each purchase is traceable and gradually shifts from what you buy to who you are. This creates the risk of AI ranking how much we are worth for links like credit, insurance or a job. Inside our own organisations, the same logic decides who is promoted, who joins the programme, and who gets labelled "high potential": all predicted by an algorithm that cannot account for the consequences. This is when privacy stops being a legal footnote and becomes an architecture of power. Without an accountable human, there is no legitimate decision. ## **Three things to take back to the office on Monday** 1. **Make "I don't know, let's test it" a leadership skill.** Maturity with AI cannot be delegated to the innovation team. Ask yourself when you last admitted, in front of your team, that you did not know something, and then went and found out with them. 2. **Design from the edge, not the centre.** Before rolling out any AI tool, process or product, ask who does not fit it yet: in your hiring, your workflows, your customer journeys. Start with them. What works at the edge usually works better for everyone. 3. **Put a name next to every inference.** List what your organisation already infers about people, customers and employees alike. For each inference, ask who decided to make it and who answers for its impact. If nobody can say, you have found your first governance task. ## **It was never about technology.** The question worth asking is not what AI will do to us. It is what we will do with AI. The Air Force did not solve its crisis with better aircraft. It solved it by admitting that the average was a fiction and redesigning around real people. We now face the same choice. The difference is that this time we have far more powerful tools, but also far fewer excuses. While we worry about the possibility of building machines that are too intelligent, the real risk lies in making decisions far too small for the world we are building with them.