Cases › made things up › LR-2026-0535

Pinsent Masons reprimanded over AI hallucinations in UK court filings

Preliminary Why: Publisher not yet named · subject to change

Status: no action reported

Severity moderate

Reported 26 May 2026

Industry legal services

No source states: loss or records, developer and model, agent tool, could do, deployer.

Severity moderate: test: data or systems of others touched with a bounded scope (one insolvency application); the firm met its former client's costs, including costs of finding new lawyers.

A junior solicitor at Pinsent Masons used an AI tool for legal research on a block transfer insolvency application in a High Court proceeding. The firm's letter to the court cited an insolvency rule that did not exist, and a follow-up letter, also prepared with AI, repeated the problem. The judge described the first letter as likely an AI hallucination that had not been checked, after querying it with the firm. The judge found the material had not been checked properly, and that supervisors had not adequately reviewed the work, although they were unaware AI had been used. The firm apologised, self-referred to the Solicitors Regulation Authority, and met its former client's costs. The judge said contempt proceedings would be disproportionate.

What happened

1 · What it was doingThe firm was instructed to prepare a block transfer insolvency application for a court matter.

2 · What went wrongA junior solicitor used AI tools for legal research, and the firm cited an insolvency rule that did not exist.

What was done

No source states an action or fix.

Risks

Made-up output relied on
Primary36 incidents

Read from the sources; not yet reviewed by a person.

Controls

Application

Not effective Human approval for high-impact actions
Did not exist User policy and training

Control failure chain

L1 · Model Not stated in sourcesProvider not named

L2 · Harness Not stated in sourcesHarness not named

L3 · Application Did not existDeployer not named

L4 · Infrastructure / target Not stated in sourcesOrganisation hit

L1 · ModelProvider not namedModel provider

No source mentions a control at this layer.

Model version not linked to a model card.

L2 · HarnessHarness not namedAgent framework vendor

No source mentions a control at this layer.

L3 · ApplicationDeployer not namedDeployer

Not effectiveHuman approval for high-impact actions“Both Cottee and Poulson checked the first draft letter going out to the court which contained the mistakes and had confirmed within around two hours that it was fine.”The Law Society Gazette

Did not existUser policy and training“It also seems to me that, as part of the supervision of LA more generally, it should have been made clear to them, if it was not made clear, that they should check all references given by AI and be candid, when reporting the result of their research to their supervisors, as to whether AI had been used in that research and whether the content generated by the AI had been checked.”The Law Society Gazette

Other controls for this risk at this layer, not mentioned in the sources: Grounding and fact checks, Output filtering and redaction, People told they are dealing with AI.

L4 · Infrastructure / targetOrganisation hit

No source mentions a control at this layer.

EffectiveNot effectiveNot effective (bypassed)Did not existNot stated in sourcesNot yet checkedAttacker's ownOrder: model · harness · application · target

Controls for this risk: result in this incident

Control	Part of the system	Result
Human approval for high-impact actions	L3 Application	Not effective
User policy and training	L3 Application	Did not exist
People told they are dealing with AI	L3 Application	Not stated in sources
Output filtering and redaction	L3 Application	Not stated in sources
Grounding and fact checks	L3 Application	Not stated in sources

Evidence

Cause a human check that existed was skipped or failed

No source states: what it reached or did, how it was caught, when it happened, when it was found, how many records were exposed, the direct loss, a count of the other parties affected, any court case, what set it off, what was done.

Show the full account: 2 items from the primary sources

What happened, in order

Not dated in the sources:

Pinsent Masons was instructed to prepare a block transfer application for an insolvency matter.
The firm has since self-referred to the SRA.

Show every field, with its source and grade

The AI system

What it was for

legal research for insolvency court proceedings

C one secondary source states it

What happened

Earliest source we hold

2026-05-26

D inferred by us from the described behaviour, reason shown

Cause

a human check that existed was skipped or failed

C one secondary source states it

Cause in detail

Supervisors did not adequately check the junior lawyer's AI-assisted research and drafts before they were sent to the court.

C one secondary source states it

What went wrong

made things up

C one secondary source states it

primary risk

RK14

D inferred by us from the described behaviour, reason shown

control chain

layer: L3; party: deployer; control id: B1; outcome: failed; quote: Both Cottee and Poulson checked the first draft letter going out to the court which contained the mistakes and had confirmed within around two hours that it was fine.; source index: 1, layer: L3; party: deployer; control id: I4; outcome: missing; quote: It also seems to me that, as part of the supervision of LA more generally, it should have been made clear to them, if it was not made clear, that they should check all references given by AI and be candid, when reporting the result of their research to their supervisors, as to whether AI had been used in that research and whether the content generated by the AI had been checked.; source index: 1

C one secondary source states it

Impact

Severity

moderate

D inferred by us from the described behaviour, reason shown

Why this severity

test: data or systems of others touched with a bounded scope (one insolvency application); the firm met its former client's costs, including costs of finding new lawyers.

D inferred by us from the described behaviour, reason shown

Context

Industry

legal services

C one secondary source states it

Notes on gaps

No dollar figures, record counts or third-party counts are stated. Resultsense claims about partner pay and other cases are excluded as unsupported by the other sources.

LR-2026-0535 · version 1 · published 2026-11-01 after automatic checks; not yet checked field by field by a person · JSON