Apple Devices Mistook Skiing Activities, Dialed False Distress Emergency Calls

Apple devices of skiers and snowboarders reportedly misclassified winter activities as accidents, which resulted in numerous false inadvertent distress calls to 911 dispatchers.

Risk classification

• Primary risk domain: 7 AI system safety, failures, & limitations
• Primary risk subdomain: 7.3 Lack of capability or robustness

The AI system failed to perform robustly under varying real-world conditions, specifically misinterpreting skiing and other high-intensity sports as vehicle crashes.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The false emergency calls were caused by the automated decision-making of Apple's crash detection algorithm, which unintentionally misclassified recreational movements as car crashes after the devices were deployed.

EU AI Act risk tier

• Risk tier: 2 High Risk

High Risk: The system is used as a safety component in consumer devices to dispatch emergency services, which has significant implications for public safety and critical infrastructure.

AI system and alleged parties

• AI system: Crash Detection (Apple)
• AI purpose: Threat Detection; Smart Devices
• Behaviour type: Autonomous
• Alleged developer: Apple
• Alleged deployer: Apple
• Alleged harmed parties: ski patrols, emergency dispatchers, Apple Watch users doing winter activities

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Minor, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Psychological

Reported: The reports describe significant frustration, stress, and operational overwhelm experienced by emergency dispatchers and responders.

Directly caused: Emergency dispatchers and search-and-rescue personnel experienced high levels of stress and frustration due to managing dozens of false alarms daily.

Inferred additional harm: Burnout and cognitive fatigue among emergency services staff in heavily affected ski regions due to the persistent influx of false alarms.

People affected

• Occurrences reported: 1
• People reportedly harmed: 50
• People reportedly exposed: 500

Potential causes

Management

• Inadequate Risk Assessment: Management released the feature without assessing impact on emergency services.

Technology

• Overly Sensitive Algorithm: Software misinterprets high-intensity recreational movements as car crashes.
• Inadequate Alert Notification: Users do not hear or feel alerts due to heavy winter clothing and helmets.
• Automatic Satellite Calling: Devices automatically route false alarms through satellite links when offline.

Data Inputs

• Misinterpreted Sensor Data: Bumping, bouncing, and rapid skiing deceleration mimic crash telemetry.
• Lack of Environmental Context: Initial software lacked GPS context to filter out non-road locations.
• Unresponsive User Input: Silence or lack of response is treated as confirmation of an emergency.

Human Factors

• Inability to Detect Alerts: Skiers wearing helmets and layers fail to notice tactile and audio warnings.
• Ignored Warning Vibrations: Some users feel vibrations but ignore them to avoid disrupting their activity.
• Poor Device Placement: Placing devices in glove boxes hides warnings and increases jarring bumps.

Process and Methods

• Inadequate Edge-Case Testing: Apple failed to test the software with extreme outdoor sports enthusiasts.
• Flawed Default Settings: Fall detection is turned on automatically for users aged 55 and older.
• Lack of Coordinated Dispatch: Dispatch centers lacked unified protocols to filter out automated ghost calls.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide consistent, detailed accounts from multiple independent emergency dispatch centers and include statements from Apple, ensuring high reliability.
• Ambiguities identified: The exact technical parameters of the crash detection algorithm are proprietary and not fully detailed.

An unintentional AI software failure in Apple devices triggered widespread false 911 calls, temporarily straining local emergency services in North American ski regions. While causing operational friction for critical infrastructure, it posed negligible threats to national security.

• Overall national security impact: Minor
• Scope: Multiple nations
• Primary target: No clear primary
• Other affected: United States, Canada
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The immediate operational strain was resolved via software updates, representing a long-term strategic lesson for automated emergency integration.
• Autonomy: Full autonomy. The crash detection algorithm autonomously initiated emergency calls to 911 without active human confirmation or intervention.
• Novelty: Evolved capability. Represents an evolved capability where consumer-wearable AI sensors autonomously interface with public safety infrastructure at scale.

Impact by dimension

• Physical security: Minor. Overwhelmed local emergency dispatchers with automated false 911 calls, temporarily straining response resources in ski regions, but managed through standard procedures.
• Information security: Negligible. No evidence of intelligence compromise, espionage, or information warfare campaigns associated with this consumer software issue.
• Sovereignty: Minor. Mildly disrupted local government emergency response operations due to automated false alarms, but did not threaten state authority or core federal decision-making.
• Economic security: Negligible. No significant threat to strategic industries, financial systems, or national technological competitive advantage.
• Societal stability: Negligible. No systemic human rights violations, mass surveillance, or widespread civil unrest resulted from the false emergency triggers.