Qualitative evidence from UK social care shows that generative-AI transcription and summarisation can introduce hallucinations, omissions and changes in wording into official records, shifting checking and accountability work onto social workers.

SocioLens project,

Claim LD05-HQ-C16 (v1.0)

Research opportunities this claim motivates

What this claim leaves open for research. Each is a hypothesis about a gap, not a finding.

Is the time AI saves paid for with weaker checking?In UK social care, AI transcription and summarisation can put hallucinations, omissions and changed wording into official records and shift checking work onto social workers.
How many AI transcription errors survive social workers' checks into case records?Qualitative evidence from UK social care shows AI transcription can introduce hallucinations, omissions and wording changes and shifts checking work onto social workers; the claim explicitly does not license a prevalence figure or a citizen-outcome effect, which is exactly what this study would measure.

Provenance and status

Claim type: Fact
Publication status: Eligible for publication, with qualifiers: emerging evidence
Verification state: Checked against the primary source. Who checked this claim is not recorded for it; claim checks so far were AI-led, and no person has checked claims against their sources.
Does not license: Population prevalence of errors; claim AI records are generally worse than human records; citizen outcome effect
Limitations: Qualitative; non-random sample; no population outcome evaluation
Source (as cited): Bruff & Groves, “Scribe and prejudice?”, Ada Lovelace Institute, 11 Feb 2026
Source type: Independent research report (recorded as “Independent qualitative report”)
Institutional position: Independent institute; vendor interviews included and labelled
Maturity of the evidence: Emerging evidence
Recorded date: 25 September 2026
Currentness / correction: This is the current version. No correction or concern is recorded on its sources.

Sources this version cites

Bruff & Groves. “Scribe and prejudice?” Ada Lovelace Institute, 11 February 2026 (Independent research report)
Read the original

Version

This is the current version, v1.0. Earlier versions stay available, and their records never change.

Versions: v1.0 · current

Raw record: every recorded field, with its coded values

All 25 canonical Claim fields, verbatim, with the coded values derived from them. The plain-language labels above are read from these.

Claim type (coded): FACT (as printed: “FACT”)
Publication status (coded): ELIGIBLE (EMERGING_EVIDENCE) (as printed: “ELIGIBLE WITH EMERGING-EVIDENCE QUALIFIER”)
Verification state (coded): CONTENT_VERIFIED_PRIMARY
Source type (coded): INDEPENDENT_RESEARCH_REPORT
Maturity (coded): EMERGING

#	Canonical field	Canonical value (verbatim)	Normalized (Claim Value Vocabulary v1.0.2)
1	CLAIM ID	LD05-HQ-C16	—
2	PUBLIC-FACING CLAIM	Qualitative evidence from UK social care shows that generative-AI transcription and summarisation can introduce hallucinations, omissions and changes in wording into official records, shifting checking and accountability work onto social workers.	—
3	CLAIM TYPE	FACT	base_types: FACT
4	RESEARCH STRAND	D — Workarounds, overrides, repair and organisational absorption	—
5	MATURITY	EMERGING	states: EMERGING
6	SOURCE	Bruff & Groves, “Scribe and prejudice?”, Ada Lovelace Institute, 11 Feb 2026	source_ids: SRC-LD05-BRUFF-GROVES-ADA-2026
7	SOURCE TYPE	Independent qualitative report	classes: INDEPENDENT_RESEARCH_REPORT
8	INSTITUTIONAL POSITION	Independent institute; vendor interviews included and labelled	—
9	SCOPE	17 local authorities in England/Scotland; interviews March–Oct 2025	—
10	DESIGN	Qualitative semi-structured interviews	—
11	SAMPLE / DATASET	39 social-care professionals/senior staff + relevant expert/vendor interviews	—
12	MODEL / TOOL / VERSION	AI transcription/summarisation tools including Magic Notes/Copilot in reported use	—
13	COMPARATOR	Prior/manual documentation practices, qualitatively	—
14	MAIN RESULT	Reported time benefits plus hallucination/misrepresentation risk; review/checking burden varies; harmful errors can enter or nearly enter official records	—
15	INFERENTIAL TARGET	Record-making as an administrative discretion/accountability site	—
16	DOES NOT LICENSE	Population prevalence of errors; claim AI records are generally worse than human records; citizen outcome effect	—
17	VERIFICATION STATE	CONTENT VERIFIED	state: CONTENT_VERIFIED_PRIMARY · verified_by: UNKNOWN
18	LIMITATIONS	Qualitative; non-random sample; no population outcome evaluation	—
19	CONTRADICTORY / SUPPORTING SOURCES	Public-sector transcription pilots; emerging evidence base	—
20	GEOGRAPHY / LANGUAGE	England/Scotland/English	—
21	DISTRIBUTIONAL LEDGER	Potential differential risk by language/accent/disability/communication context; outcome evidence incomplete	—
22	R01 / FIELD-MAP LINKS	D5, D2, D3, L-C	—
23	WHAT WOULD CHANGE THIS CLAIM?	Large comparative study finding negligible record effects or robust automated correction	—
24	PUBLICATION STATUS	ELIGIBLE WITH EMERGING-EVIDENCE QUALIFIER	state: ELIGIBLE · qualifiers: EMERGING_EVIDENCE
25	RECORDED DATE	2026-09-25	—

Field 22 references resolved against Field Map v0.2: D5 (DOMAIN); D2 (DOMAIN); D3 (DOMAIN); L-C (LENS)

The sources as recorded, with their record IDs and coded source types:

Pinned Sources (field 6, version-specific)

“Bruff & Groves, “Scribe and prejudice?”, Ada Lovelace Institute, 11 Feb 2026” → Bruff & Groves. “Scribe and prejudice?” Ada Lovelace… (SRC-LD05-BRUFF-GROVES-ADA-2026--v1.0) — Bruff & Groves. “Scribe and prejudice?” Ada Lovelace Institute, 11 February 2026; class INDEPENDENT_RESEARCH_REPORT (explicit public provenance pin)

Cited by Living Debate versions

Discretion curtailed, relocated or encoded? (v0.1) LD-05

Cite this page

SocioLens project. (2026, September 25). Qualitative evidence from UK social care shows that generative-AI transcription and summarisation can introduce hallucinations, omissions and changes in wording into official records, shifting checking and accountability work onto social workers [Claim LD05-HQ-C16, version v1.0]. SocioLens. https://sociolens-web.pages.dev/claims/ld05-hq-c16/v1.0/

@misc{sociolens_ld05_hq_c16_v1_0,
  author = {{SocioLens project}},
  title = {Qualitative evidence from UK social care shows that generative-AI transcription and summarisation can introduce hallucinations, omissions and changes in wording into official records, shifting checking and accountability work onto social workers.},
  year = {2026},
  date = {2026-09-25},
  howpublished = {\url{https://sociolens-web.pages.dev/claims/ld05-hq-c16/v1.0/}},
  note = {Claim LD05-HQ-C16, version v1.0}
}

The link points to this version, whose record never changes. For a finding, cite the original study as well: SocioLens summarizes research and does not replace it.

Research-of-record provenance

Materialized by sociolens-materializer 0.6.0 at 2026-09-28T20:00:00.000Z; normalization contract SocioLens — Claim Value Vocabulary v1.0.2 — Pre-Spike Materialization Contract (v1.0.2). Content seal SHA-256 2e5e4ddcdfb933a6e3d9db3fb9d5a7383f1b57198838a43a3b2bd5a4734fa045.

Frozen Drive-derived inputs

R04 — LD-05 Canonical Claim Ledger v1.0 — LOCKED (v1.0) — Drive ID 1WkAv7jmWIrr1GBmYSv66jDc8m222Yd9z; revision: NOT_EXPOSED_BY_DRIVE_TEXT_EXPORT; Drive modified ; block “ENTRY 16 — LD05-HQ-C16” SHA-256 2e41af013a90ca7f4daa816cb35f6ec25dae1602ab42a8a68ea545dcdc42ffa5; SocioLens — Programme Status & Authority Register v0.6 status “CURRENT.”; frozen fixture fixtures/frozen/r04-ld05-claim-ledger-v1.0.

Builder decisions (not Drive data)

VERSION = v1.0 — basis: Claim version taken from the canonical ledger artifact version (R04 — LD-05 Canonical Claim Ledger v1.0); Implementation Architecture Selection §5
CURRENT_POINTER = true — basis: SocioLens — Programme Status & Authority Register v0.6 status for "R04 — LD-05 Canonical Claim Ledger v1.0 — LOCKED": "CURRENT."
FIELD_6_SOURCE_REFERENCE = SRC-LD05-BRUFF-GROVES-ADA-2026--v1.0 — basis: field 6 of ENTRY 16 — LD05-HQ-C16 cites "Bruff & Groves, “Scribe and prejudice?”, Ada Lovelace Institute, 11 Feb 2026"; the frozen field-6 segment of this ClaimVersion's ledger block; Drive-derived identity authority (Issue #37 Source inventory): ledger C16 fields 6–8; Batch 03 §8; KEY EVIDENCE.
SOURCE_PROVENANCE_PIN = SRC-LD05-BRUFF-GROVES-ADA-2026--v1.0 — basis: Issue #5 proof rendering (Architecture Proof Protocol v0.1 AP-08 / Amendment v1.2 §6): explicit version-pinned public provenance of this ClaimVersion — the pinned field-6 Source record.