THE QUESTION WE STARTED WITH
Who gets a better life when AI gets better?
Model capability and practical adoption move at different speeds. A faster task becomes a useful service when the organisation can show who benefits and who checks mistakes.
THE NEXT QUESTION
In one Greek service, who actually receives the time that AI saves, and what would prove that their life or work improved?
Follow one Greek service from the employee doing the work to the citizen receiving it. Check whether speed, accuracy, access and the ability to correct a mistake improve together.
Organizing a discussion on this question?
Use these findings and sources as the starting point for a panel, interview, email exchange or small roundtable.
- 01The starting point
Model capability and practical adoption move at different speeds. A faster task becomes a useful service when the organisation can show who benefits and who checks mistakes.
- 02Further evidence
A customer support study found a 15% average productivity gain, while METR’s early in 2025 developer trial found a slowdown. METR’s February 2026 update warns that selection and concurrent agents make newer estimates unreliable. The practical comparison must include complete cases, checking and omitted work.
Still open These settings and tools do not predict a Greek service’s result. No local trial or universal productivity effect is established.
- 03The decision to test
Looking at the whole job raises a harder question: who gets the time AI saves? Any decision to expand the service must count quality, difficult cases and all staff effort. It must also say which result would make us stop.
Still open The actual gain and its distribution remain unknown. METR’s methodological revision reinforces the need to record selection, rather than treat a headline estimate as a local target.
WHO CAN HELP ANSWER THIS?
The people closest to the problem.
Frontline employee
Show one task before and after AI. Which work disappeared and which new work appeared?
Comparable task records, time spent checking and correcting output, and escalations.Service user, including someone who needs accessibility support
Was the service easier to complete, and could you get a person to resolve a mistake?
Completion time, repeat contacts, accessibility barriers and unresolved complaints.Service owner or SME lead
What would make you expand, redesign or stop this deployment?
Original objective, baseline, total operating cost and a named decision owner.Each perspective can fill a different gap in what we know.
Follow the question across the network.
ThoughtLeaders.gr The attributed positions ↗AIStrategy.gr The question in this sector ↗EventChanger The conversation ↗Sources and what remains uncertain
Read the original work alongside the analysis. For the local case, we still need the evidence requested above.
- Dario AmodeiWe are near the end of the exponential ↗13 February 2026 · amodei-dwarkesh-20260213
- Satya NadellaHow Microsoft is preparing for AGI ↗12 November 2025 · nadella-dwarkesh-20251112
- Erik Brynjolfsson, Danielle Li, Lindsey RaymondGenerative AI at Work ↗4 February 2025 · qje-ai-at-work-20250204
- Joel Becker, Nate Rush, Beth Barnes, David ReinMeasuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity ↗10 July 2025 · metr-developer-productivity-20250710
- Joel Becker, Nate Rush, Tom Cunningham, David Rein, Khalid MahamudWe are Changing our Developer Productivity Experiment Design ↗24 February 2026 · metr-uplift-design-20260224