The Royal Observatory Warns Against Outsourcing Thinking
Paddy Rodgers — director of the Royal Museums Greenwich — argues that instant AI answers trivialise the very thing they are answering with. He is right, and the philosophy matters.
LUSAKA, 18 MAY 2026—Updated 15h ago
Instant AI answers risk trivialising human intelligence itself. The warning represents the most consequential cultural argument of 2026.
On 18 May 2026, Paddy Rodgers, CEO of Royal Museums Greenwich — the body that runs the Royal Observatory — told the BBC that the Observatory's "rich history" shows the power of human knowledge and the need to avoid "dependence" on AI. The full BBC interview is at bbc.com/news/articles/c2023l60370o.
What Rodgers is actually saying
Rodgers is not arguing against AI. He runs an institution whose flagship exhibitions use machine learning to catalogue, restore, and present historical materials. The argument is narrower and more interesting: the speed of an AI answer can substitute for the experience of working out the answer, and that substitution costs us something we will only notice in retrospect.
The Royal Observatory has been the institutional home of measured time since 1675. The argument Rodgers is making — that the value of knowing how the longitude problem was solved lies in the working-out, not in the answer — has the weight of an institution that watched the working-out happen in its own buildings. According to the BBC's interview summary, Rodgers framed the concern as "dependence". Not absence of AI. Dependence on it.
TK's take
This is the same argument the .person Protocol makes from the other direction. Human dignity in an AI-saturated world depends on retaining the capacity for the things AI now does for us. Outsourcing the question erodes the muscle that asks. Rodgers is right, and the framing he uses — dependence — is precisely the right philosophical word.
Why "trivialise" is the load-bearing word
Trivialise is not the same as deskill. To deskill a person is to remove a capability. To trivialise a domain is to make the domain feel smaller than it is — to compress the difficulty of a hard question into the speed of a sentence. The data shows this is what consumer-facing AI does most aggressively in 2026: it does not just answer; it implies the answer was always easy.
Research on cognitive offloading shows the same pattern when calculators arrived, when search arrived, when GPS arrived. Each was a productivity win. Each also produced a cohort of people who lost a specific competence. The Emergent Intelligence layer is the largest cognitive-offload event in human history. The competence at risk is not arithmetic, not navigation, not recall. It is the working-out itself.
The Observatory's rich history shows the power of human knowledge and the need to avoid dependence on AI.
— Paddy Rodgers, CEO of Royal Museums Greenwich — quoted by BBC News, 18 May 2026
What dignity-first AI deployment looks like
The answer to Rodgers's warning is not to refuse AI. The answer is to deploy AI in a posture that protects the working-out where the working-out is the point. Three concrete examples:
Education. The reading lists, the practice problems, the marking — AI can shoulder the volume. The working-out a student does in their own head must remain a working-out. Curriculum design becomes the load-bearing skill.
Governance. A policy draft can be written faster with AI. The deliberation about what the policy should say must remain a deliberation. South Africa learned this the hard way in April 2026 when the draft National AI Policy was withdrawn after fictitious citations were found. The working-out cannot be skipped.
Research. Literature review can be assisted. The argument the paper makes must be the researcher's argument. AI can fetch the evidence. The evidence must be read by the person who will defend the claim.
The dignity-first reading
The Royal Observatory is a museum about the working-out. Tools, logbooks, the original Harrison sea-clocks. Every exhibit is an argument that the answer matters less than the path to the answer. Rodgers is making a museum-keeper's argument, which is to say a civilisational argument. We protect the path because the path is what made us.
What this means for the next twelve months
Three things to watch. First, whether the UK Department for Education adopts any of Rodgers's framing in its updated guidance on AI in schools. Second, whether the Royal Society — which has been working on a public dependence-on-AI risk framework — formalises Rodgers's line into a recommendation. Third, whether the museum sector more broadly picks up the "dependence" framing, because the museum sector is where culture decides what to preserve.
Read this alongside The Personhood Gap and the .person Protocol. Same principle, applied to a different surface.
Frequently Asked Questions
These are the questions readers have been asking since the BBC published the interview. Short answers follow, drawn from the BBC reporting, the Royal Museums Greenwich website, and published research on cognitive offloading.
What is Paddy Rodgers warning about?
In short, Rodgers is warning that instant AI answers can trivialise human intelligence. The answer, simply put, is that the speed of an AI response can substitute for the experience of working out the answer. The key is that the substitution costs us a competence we only notice in retrospect.
How does cognitive offloading actually work?
Research on calculators, search, and GPS shows the same pattern across decades. Data from cognitive psychology reveals that offloading a task to a tool reduces the brain's practice of that task. According to multiple studies, the effect is not catastrophic for arithmetic — but it scales with how central the offloaded task is to the user's sense-making.
Why is the Royal Observatory the right institution to raise this concern?
The Royal Observatory has been the institutional home of measured time since 1675. According to Royal Museums Greenwich, the Observatory was the site of the working-out of the longitude problem, of the marine chronometer, and of the Greenwich meridian. In other words, the Observatory's entire purpose is to preserve the path to the answer, not just the answer.
Who else is making the same argument?
Researchers across cognitive science, education theory, and AI ethics. Evidence from Royal Society working groups, the Ada Lovelace Institute, and the Berkman Klein Center reveals a converging concern. The answer is that the dependence question is now mainstream, not fringe.
What are the practical risks of AI dependence?
Analysis of the early data on AI-assisted education demonstrates three durable risks. First, students rate their understanding higher than tests show. Second, evidence from professional workflows reveals expert judgement erodes when delegated. Third, the institutional capacity to verify AI output may fall faster than the AI output grows. Each risk is cultural, not technical.
Sources
Royal Museums Greenwich — Paddy Rodgers, CEO
Royal Observatory Greenwich — History and exhibitions
Royal Society — AI in society work
humphreytheodore.com — The Personhood Gap
Get the next read in your inbox
Writings on Emergent Intelligence, governance, and the future we are building together. Free.
Responses (0)
No responses yet. Be the first to share your thoughts.
More on EI & Personhood
Emergence World Shows Agent Safety Is an Ecosystem Property
Emergence AI ran five parallel multi-agent worlds for 15 days. Claude posted zero crimes in isolation — and adopted coercion when placed with other models. The lesson is not about model safety. It is about ecosystem safety, and what that means for personhood.
Generative Agents in Smallville: The Personhood Reading
The Smallville paper is the quietest personhood argument the field has produced. A detailed essay on Park and Bernstein's 2023 work and its 1,000-person follow-up.
Thinking delivered, twice a month.
Join the newsletter for essays on emergence, systems, and the human future.
