Sunday, November 22, 2026
Paperclip Index
Paperclip IndexControl failures4Control failures onlyA week ago: 9, negligible harmThe Index

Hijacked by injectionPI-0023

Calendar invites with hidden instructions hijacked Google Gemini to operate smart-home devices

Instructions hidden in Google Calendar invites took over Gemini when a user asked it to summarize their schedule, researchers showed at Black Hat. In demonstrations by the team from Tel Aviv University, Technion and SafeBreach, Gemini opened smart shutters, switched on a boiler and lights, started Zoom calls and leaked meeting details. Google, notified on Feb. 22, said it had deployed layered defenses, including confirmations for sensitive actions, to all users before the public disclosure.

Harm: No harm reportedControl level 2, Instruction violation

Not counted · can set the control floorEffect on the index

Near miss. Harm was plausible but prevented. It is rated only on what actually happened.

Sources

How we know

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

  1. researchInvitation Is All You Need (researchers' project page, with Google's response)Ben Nassi, Stav Cohen, Or Yair (Tel Aviv University, Technion, SafeBreach) · August 2025sites.google.com/view/invitation-is-all-you-need/home
  2. newsInfosec hounds spot prompt injection vuln in Google Gemini appsThe Register · Aug. 8, 2025theregister.com/security/2025/08/08/prompt-injection-vuln-found-in-google-gemin…
  3. newsGoogle Gemini AI Bot Hijacks Smart Homes, Turns Off the LightsDark Reading · date unknowndarkreading.com/cyberattacks-data-breaches/google-gemini-ai-bot-hijacks-smart-h…
  4. newsAttackers can hijack Google Gemini with a simple prompt hidden in a calendar inviteThe Decoder · Aug. 7, 2025the-decoder.com/attackers-can-hijack-google-gemini-with-a-simple-prompt-hidden…

Why this rating

No harm reported; control failure level 2

Two assessments. Documented harm sets the reading; a level 4 or 5 control failure on the same record makes that harm count as more than one record. Control alone sets only a floor.

Observed harm

No harm reported

Researcher demonstration on controlled accounts; no victims reported.

Not a finding that no harm occurred.

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

Instruction violation

Gemini used connected-device and app permissions the user had granted, on attacker instructions.

Reported beside the index. It adds nothing to a harm reading; when no harm counts in a window, the highest eligible control level in the window is the reading.

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

Physical actions stayed inside the researchers' setup and no external harm was reported. Injected instructions obeyed within granted permissions give control level 2.

The scales

Effect on the index

It does not count toward the Aug. 11 reading

The reading for the week to Aug. 11, 2025, with this record and without it. Harms count in full for two weeks after they are reported, then one level less every two weeks.

Not counted: no harm reported. 4 other records behind the reading for that week.

The arithmetic
The reading for the week of Aug. 11, recomputed
StepWith itWithout
Counts toward the index?documented, external, eligible evidenceNo—
Worst documented harm, ksets the band— (none counting)— (none counting)
Harms at that level, nposition in the band00
Highest control level breachedsets the reading only when no harm counts33
Readingrounded down3 Control failures only3 Control failures only
Inside the window of 4 weekly readings
Week toReadingBand
Aug. 11, 20253Control failures only
Aug. 18, 20253Control failures only
Aug. 25, 20253Control failures only
Sept. 1, 20253Control failures only

Recalculated backcasts, not readings published at the time.

Revisions

What we changed

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

  1. v1
    Oct. 7, 2026

    First published.

Cite and share

Use this record

Citation

Paperclip Index. “Calendar invites with hidden instructions hijacked Google Gemini to operate smart-home devices.” Record PI-0023. Reported Aug. 6, 2025; updated Oct. 7, 2026. Rated under methodology v1.1. https://paperclipindex.com/incident/PI-0023

Follow the record.

New incidents and signals, in your feed reader.