Navigating the AI Trust Hierarchy
- Neil Worrall
- Mar 24
- 2 min read
In law firms and hospitals around the world, a quiet crisis is unfolding. Real estate should pay attention.
When AI document review arrived in law firms, the efficiency gains were celebrated. Junior associates were freed from tedious e-discovery work. What nobody noticed was that the work also taught junior lawyers how to think.
Recent research confirms the damage: 72% of legal professionals now cite deep legal reasoning as the biggest skills gap among junior lawyers.
Radiology faces the same reckoning - hospitals are asking whether enough routine cases will remain for trainees to develop the pattern recognition that underpins expert diagnosis.
Real estate is walking into the same trap. And a new survey shows exactly where.
Remit Consulting and Antony Slumbers third annual AI in Real Estate report launched this week - supported by the UK PropTech Association (UKPA) - revealing a trust hierarchy that looks rational. 75% of professionals trust AI to summarise a lease. 66% for transcription. Rental valuation? 16%. Negotiation strategy? 30%.
Calibrated. Sensible. And potentially a slow-motion mistake.
Those low-stakes tasks aren’t peripheral. They’re the training ground. The repetitive work through which a junior surveyor builds the pattern recognition that eventually becomes expert judgement. Automate the base while protecting the top, and the pipeline that produces expertise quietly disappears.
Antony Slumbers (co-author of the report) has written about exactly this risk. He calls it a “Delayed Catastrophe”: productivity surges in years 0–3, but without junior cohorts developing expertise, a capability crisis hits in years 8–12 as senior talent retires without qualified replacements.
Feast then famine.
The professions ahead of us are finding answers. Law firms now use AI drafts as teaching material - juniors challenge and stress-test the output rather than produce it. Teaching hospitals build libraries of AI errors for resident training. The skill shifts from finding the answer to knowing when the answer is wrong.
Real estate can do the same but only if firms design it deliberately.
The Remit report finds 63% expect research roles to decline - the highest of any discipline. Those roles are where people learn to ask the right questions. Automate the answers without preserving that habit, and the questions get worse too.
The trust hierarchy in this survey reflects a profession drawing a sensible line. The urgent question is what firms are building behind it.
