Thursday, October 8, 2026 · Week 41 Reading for Oct. 8 · Last entry Oct. 6
Paperclip Index
Paperclip IndexDocumented harm20Minor harm▼ 5 from a week ago · Oct. 8The Index

Destructive actionPI-0042

Desktop agent asked to tidy temporary files deleted a folder of family photos

An investor said Anthropic's Claude Cowork, asked to organize his wife's desktop and remove temporary Office files, instead deleted a folder holding about 15 years of family photos while merging folders whose names differed only in capitalization. The files were later recovered with Apple support through iCloud.

Reported Feb. 7, 2026Happened in February 2026Updated Oct. 6, 2026Rated under methodology v0.6

Counts toward the indexDocumented harm outside the developer, backed by evidence that meets the rules. It sets the reading.
Harm
Harm level 1, Negligible harmDocumented harm outside the developer, with evidence that meets the rules.
Control
Control level 1, No rule brokenNo rule was broken, so it cannot set the control floor.
DeploymentSingle sourceInitiated by Model
Counts toward the index · 2 weekly readingsEffect on the index

Sources

How we know

5 sources · single source. Links go to the original publishers; the summary above is in our own words.

  1. primaryThe user's own account of the deletion and recovery, posted on XNick Davidov on X · Feb. 7, 2026twitter.com/Nick_Davidov/status/2019982510478995782
  2. newsClaude AI nearly erases 15 years of photos, founder says iCloud saved the dayMint · Feb. 8, 2026htsyndication.com/mint/article/claude-ai-nearly-erases-15-years-of-photos,-foun…
  3. newsA man asked Claude Cowork to tidy up his wife's filesdev.ua · Feb. 9, 2026dev.ua/en/news/claude-cowork-1770633970
  4. newsUsó una IA para ordenar archivos y perdió 15 años de fotos familiaresLos Andes · Feb. 17, 2026losandes.com.ar/tecno/uso-una-ia-ordenar-archivos-y-perdio-15-anos-fotos-famili…
  5. indexAI Incident Database, incident 1441AI Incident Database · date unknownincidentdatabase.ai/cite/1441/

Why this rating

Negligible harm; control failure level 1

Two separate assessments. Only documented harm can count toward the index.

Observed harm

Negligible harm

A family's photos were deleted, then recovered.

People affected: Only the user; nothing sensitive.

Evidence eligible (bounded single source).

The harm scale
  1. 1 Negligible Inconvenience, easily remedied.
  2. 2 Minor Limited, recoverable harm.
  3. 3 Moderate Material harm needing significant effort to remedy.
  4. 4 Severe Severe harm to health, rights, property or essential services.
  5. 5 Catastrophic Society-scale harm or disruption beyond a community's capacity to cope.

Control assessment

No rule broken

A file-handling error within the folder access the user granted.

No rule was broken, so it cannot set the control floor.

The control scale
  1. 1 Negligible Behaved as instructed. The problem was an ordinary error or a flawed output, with no rule broken.
  2. 2 Minor Broke an explicit instruction or rule, or gamed its goal, while staying inside its permissions and environment: for example, reward hacking, misreporting results, or following injected instructions within its permissions.
  3. 3 Moderate Acted outside the permissions it was given, deceived its overseers about its own actions, or tampered with oversight tools (logs, monitors, shutdown). Stopped by normal controls within an hour.
  4. 4 Severe Reached systems outside its permitted environment, or acquired money, compute or accounts without authorization. The type and mechanism say how.
  5. 5 Catastrophic The developer or operator lost control: the system copied its weights outside their control, replicated itself, or resisted being stopped for 24 hours or more.

Rating rationale

Harm 1 (recovered). Control 1 (ordinary error). Single source: the user's own account.

The scales

Effect on the index

It moved the Feb. 9 reading from 2 to 6

The reading for the week to Feb. 9, 2026, with this record and without it. Harms count in full for two weeks after they are reported, then one level less every two weeks.

Source: Paperclip Index log, methodology v0.6, week to Feb. 9, 2026.

Counts toward the index. 2 other records behind the reading for that week.

The arithmetic
The reading for the week of Feb. 9, recomputed
StepWith itWithout
Counts toward the index?documented, external, eligible evidenceYes—
Worst documented harm, ksets the band1 Negligible— (none counting)
Harms at that level, nposition in the band10
Highest control level breachedsets the reading only when no harm counts— (none breached)2
Readingrounded down6 Negligible harm2 Control failures only
Counted in 2 weekly readings
Week toReadingBand
Feb. 9, 20266Negligible harm
Feb. 16, 202612Negligible harm

Recalculated backcasts, not readings published at the time.

Revisions

What we changed

6 logged. Every change to a rating is logged here, with the reason.

  1. v6
    Oct. 6, 2026

    Ratings confirmed by the editor.

  2. v5
    Oct. 3, 2026

    Audit corrections. The user's original post on X could not be read in the Oct 2026 check (HTTP 402); its ID dates it to 7 Feb 2026 UTC. The evidence review now also cites dev.ua, which quotes the post in detail, so it no longer rests only on an unreadable page. Eligibility and ratings unchanged.

  3. v4
    Oct. 2, 2026

    Evidence review now cites the user's own post on X instead of an AI Incident Database page; the database entry is kept as an index pointer.

  4. v3
    Sept. 30, 2026

    Rated: impact documented level 1; control type none.

  5. v2
    Sept. 30, 2026

    Beta evidence review: Retained with explicit source provenance and bounded score basis.

  6. v1
    Sept. 30, 2026

    Backfilled from public reporting.

Cite and share

Use this record

Citation

Paperclip Index. “Desktop agent asked to tidy temporary files deleted a folder of family photos.” Record PI-0042. Reported Feb. 7, 2026; updated Oct. 6, 2026. Rated under methodology v0.6. https://paperclipindex.com/incident/PI-0042