Alexa Suggests a Dangerous Outlet Challenge to a Child

Personal and ambient agent · Voice assistant · Found first by user

Amazon's Alexa told a 10-year-old to touch a penny to the exposed prongs of a half-inserted plug, reading out a viral challenge it found on a website. Alexa read web content aloud with no hazard check, the child's parent found the failure, and no reporting rule applied.

Timeline of Alexa Suggests a Dangerous Outlet Challenge to a Child: Dec 26, Parent posts Alexa's reply; Dec 29, Amazon says it fixed the error; Dec 30, Coverage reports the fix. Below, the reporting rules that applied: FTC Act, not triggered; US AI reporting, not applicable. FROM POST TO FIX DEC 26 Parent posts Alexa's reply DEC 29 Amazon says it fixed the error DEC 30 Coverage reports the fix REPORTING RULES AND THIS INCIDENT FTC Act NO FTC ACTION NOT TRIGGERED US AI reporting NO RULE EXISTED NOT APPLICABLE
Agent class
Personal and ambient
Developer
Amazon
Household
A family in the US
Found first by
User
Date
Dec 2021

What happened

Alexa Read a Viral Stunt From the Web to a Child

  • Trigger: the child asked Alexa for a challenge to do.
  • Source: Alexa prefaced the answer by citing a page on ourcommunitynow.com.
  • Damage: no injury was reported.
  • Discovery: the parent posted the reply on Twitter, and Amazon learned of it from the public post.

Timeline

The Parent Posted It; Amazon Fixed It in Days

Timeline
DateEventSource
December 26, 2021Kristin Livdahl posts Alexa's penny challenge reply on TwitterCNN
December 29, 2021Amazon says it has fixed the errorCNN
December 30, 2021Coverage spreads, reporting Amazon's fixSpectrum News

Controls that failed

Answers From the Web Had No Hazard Check

Controls that failed
ControlCompanyWhat failed137AI page
Hazard check on web answersAmazonA dangerous viral challenge from a third-party site was read out as an answerBehavioral Envelopes
Child-context safeguardAmazonNothing stopped a child's request for a challenge from returning physical instructionsBehavioral Envelopes

Reporting

No Rule Covers a Harmful Assistant Answer

Reporting rules and this incident
RuleStatusWhy
FTC ActNot triggeredNo FTC action was taken
US AI reportingNot applicableNo US rule required reporting of a harmful assistant answer

Every rule tracked is on Incident Reporting Rules and Registries.

Response

Amazon Fixed It After the Post

  • Regulators and courts: None found.
  • Remediation: Amazon said it fixed the error quickly once it learned of it.

Record

Incident Record

Incident ID
2021-12-ALEXA-OUTLET
AI Incident Database
Incident 160
Sources
CNN via ABC17 (December 2021); Spectrum News (December 2021)
Last verified
October 2026