Experiment
DREAM BIG
WHAT IF
Defra’s understanding of land kept itself aligned with what is happening on the ground?
Living Land Record
Keep Defra’s understanding of land aligned with what is happening on the ground.
01
Problem to address
Defra services depend on an accurate understanding of land, but the real world changes continuously while administrative records are updated through reporting, inspections, mapping jobs and hand-offs. When the record lags behind reality, farmers repeat information, field officers spend time verifying known change, and downstream decisions can be made using stale evidence.
02
What the evidence tells us
Land change is often discovered during visits, marked as a land information discrepancy, returned through Siti Control and then passed into a separate mapping/digitising process before records catch up. People repeatedly describe stale or conflicting records, duplicated correction and the need to establish a trustworthy current picture. This concept is a DREAM BIG synthesis of those evidenced problems, not evidence that automated land-change detection is already proven.
03
The proposition to explore
The idea
Create a living land record that continuously compares Defra’s current understanding with new evidence from the real world, surfaces likely changes, and asks farmers or staff to verify only where human confirmation or judgement is genuinely needed.
04
Hypothesis
If Defra can detect likely land changes earlier and route only uncertain changes for verification, its understanding of land will stay closer to reality while reducing repeated reporting, avoidable mapping hand-offs and unnecessary field verification.
05
How it could work
Combine the authoritative land record with appropriate new signals such as satellite or aerial imagery, previous inspections, farmer updates and other verified spatial evidence. Detect meaningful change, assign a confidence level and determine the next action. High-confidence changes might trigger a lightweight confirmation; ambiguous or consequential changes would be routed to the right person for review. Once verified, the change should propagate to services that depend on it rather than being corrected separately in each place.
06
Example: bring it to life
A hedge line and field boundary appear to have changed since the last verified record. Before the next inspection, the service surfaces the change to the farmer: “We think Parcel 42 may have changed since March. This could affect your current agreement. Has something changed?” The farmer confirms one change and disputes another. When the field officer later visits, they see only the disputed area requiring verification rather than rediscovering all of the change from scratch.
07
AI opportunity
AI or computer vision could help detect spatial change, compare evidence and explain why a change was surfaced. It is not the whole proposition. Authoritative identifiers, spatial data, provenance, confidence rules and governed update workflows are more fundamental. High-impact changes should not be silently accepted by an opaque model.
08
Potential value
- More accurate and current land evidence
- Less repeated reporting and manual reconciliation
- More targeted field visits and mapping work
- Fewer downstream decisions based on stale land information
- A stronger foundation for personalised interventions, outcome modelling and proactive funding
09
Reduce uncertainty without overcommitting
Smallest meaningful experiment
Take one geography and one clearly observable type of land change. Use historical imagery and known verified changes to test whether candidate changes could have been surfaced earlier with useful confidence. Then prototype the verification experience with farmers, field officers and mapping staff to see which changes could be confirmed remotely and which still require expert review.
10
Open questions and risks
What we still need to learn
- Not every meaningful land change is visible remotely
- False positives could create more work rather than less
- Legal and policy rules may require explicit notification for some changes
- Defra must be clear about which record is authoritative and how disputes are resolved
- Continuous observation raises trust, transparency and governance questions