Starship Technologies Delivery Robot Reportedly Injures Arizona State University Employee

A semi-autonomous delivery robot operated by Starship Technologies reportedly struck a pedestrian employed by Arizona State University on the campus sometime in September 2023, purportedly causing injuries after abruptly reversing into her. The robot is reported to have initially knocked the pedestrian down, then moved toward her again while she was still on the ground. The company reportedly offered the victim promo codes and insurance information as an apology. On September 19, 2024, 404 Media made the police report of the incident available.

Risk classification

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

The incident was caused by a navigation failure where the robot unexpectedly reversed into a pedestrian twice, demonstrating a lack of capability and robustness in dynamic pedestrian environments.

Additional risk subdomains

• 1.1 Unfair discrimination and misrepresentation: The company requested that promotional photos only feature students with a healthy-looking body mass index, representing discriminatory representation practices.

Causal factors

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

The physical harm was caused by the autonomous navigation decisions of the deployed Starship delivery robot, which unexpectedly reversed into a pedestrian.

EU AI Act risk tier

• Risk tier: 2 High Risk

High Risk: The system is an autonomous physical robot operating in public pedestrian infrastructure, where navigation failures present direct physical safety risks to the public.

AI system and alleged parties

• AI system: Starship delivery robot (Starship Technologies)
• AI purpose: Autonomous Driving; Navigation Assistant
• Behaviour type: Autonomous
• Alleged developer: Starship Technologies
• Alleged deployer: Starship Technologies
• Alleged harmed parties: Unnamed Arizona State University employee

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 Minor, 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 Minor, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Physical

Reported: The report explicitly describes physical injuries to the ASU parking employee caused by the robot collision.

Directly caused: The victim suffered lower back pain, a 4-inch cut on her left arm, and had difficulty walking after being knocked down twice by the robot.

Differential treatment

Reported: The report explicitly describes a corporate policy requiring differential treatment of individuals based on physical appearance.

Directly caused: Starship requested that universities only take promotional photos with students who have a healthy-looking body mass index.

People affected

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

Potential causes

Management

• Prioritizing PR Over Safety: Focus on promotional photos and promo codes rather than safety fixes.
• Inconsistent Legal Actions: Unclear policies on when to prosecute or cooperate with authorities.

Technology

• Erratic Reversing Behavior: Robot abruptly reversed and struck a pedestrian twice while on the ground.
• False Obstacle Detection: System claimed to detect an oncoming vehicle when none was present.
• Navigation Failures: Robots drove into construction zones and off campus boundaries.

Data Inputs

• Spurious Sensor Readings: Robot detected a non-existent oncoming vehicle, triggering a reverse.
• Blind Spot in Rear Sensors: Robot failed to detect the pedestrian standing directly behind it.

Human Factors

• Pedestrian Misjudging Robot: Pedestrian crossed behind the robot, assuming it would continue forward.
• Remote Operator Delay: Robots got stuck at traffic lights waiting for remote human assistance.

Process and Methods

• Inadequate Incident Reporting: Company was reluctant to share video footage with affected parties.
• Flawed Investigation Protocol: Company insisted on leading investigations rather than independent police.
• Inadequate Operational Boundaries: Lack of robust geofencing allowed robots into construction zones.

Regulatory Environment

• Lack of Autonomous Vehicle Laws: No clear local regulations governed sidewalk delivery robot operations.
• Inadequate Accountability Standards: No regulatory framework forced the company to share sensor/video data.

Information quality

• Classification confidence: High
• Reason for confidence: The primary incident is well-documented through official police reports and corroborated by video footage reviewed by law enforcement, providing high certainty about the sequence of events.
• Ambiguities identified: There is a conflict between Starship's claim that the robot reversed to avoid an oncoming vehicle and the police officer's observation of the video footage showing no vehicle or obstacle was present.

A localized incident where a commercial autonomous delivery robot malfunctioned and struck a pedestrian on a university campus. The incident has negligible national security implications, representing a standard commercial safety and navigation failure rather than a threat to national infrastructure, sovereignty, or security.

• Overall national security impact: Negligible
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The event is a past, localized incident with no ongoing threat or active crisis requiring urgent national response.
• Autonomy: Full autonomy. The delivery robot operates and navigates public environments autonomously without active human intervention or direct real-time control.
• Novelty: Established threat. Minor navigation failures and low-speed collisions involving autonomous ground vehicles are established risks with existing precedents.

Impact by dimension

• Physical security: Negligible. The incident is a localized, low-speed collision involving a commercial delivery robot and a pedestrian, posing no threat to critical infrastructure or national physical security.
• Information security: Negligible. No intelligence compromise, classified data theft, or information warfare activities were involved in this localized commercial navigation failure.
• Sovereignty: Negligible. The incident occurred on a university campus and does not impact state authority, electoral processes, or core government operations.
• Economic security: Negligible. The failure of a commercial food delivery robot does not threaten strategic industries, critical supply chains, or national technological competitiveness.
• Societal stability: Negligible. While the company had a discriminatory PR policy regarding body mass index, the incident does not pose a threat to large-scale societal stability or systemic human rights.