This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
SoftBank's Pepper robot, marketed as a sophisticated humanoid assistant, failed to meet performance expectations across various sectors including funeral services, nursing homes, and banking. Users reported frequent mechanical breakdowns, inability to recognize faces, and limited functionality, resulting in contract terminations and financial losses for businesses and individuals. The product line was eventually discontinued as simpler, more reliable alternatives like smart speakers proved more effective.
SoftBank's robot allegedly kept making mechanical errors, taking unplanned breaks, failing to recognize previously-met people, and breaking down during practice runs.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
Also classified as
Users anthropomorphized the humanoid robot and expected human-level intelligence, leading to disappointment and financial investment based on unrealistic expectations.
The Pepper robots repeatedly failed to perform their tasks reliably, experiencing mechanical errors, face recognition failures, and unexpected shutdowns during operation.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by the Pepper robot failing to perform its programmed tasks post-deployment due to technical limitations and mechanical errors.
The failure and eventual discontinuation of SoftBank's Pepper robot represents a commercial product failure with negligible national security implications. The issues were limited to technical and mechanical limitations in consumer and business assistant roles, causing minor financial losses for users but posing no threat to critical infrastructure, sovereignty, or national stability.
The report describes Pepper as a humanoid assistant designed to augment human capabilities in nursing homes, bank lobbies, and homes by greeting guests, leading exercises, and recommending products, while requiring human oversight.
Risk Level 3: Limited Risk. Pepper is a humanoid robot designed to engage in basic conversation and interact with humans, which functions as a physical chatbot and thus falls under the transparency requirements of Limited Risk systems.
Pepper was primarily designed to interact with humans through conversation, facial recognition, and basic dialogue, functioning as a physical chatbot. It was also used to recommend financial products to bank customers.
The reports provide clear, consistent, and detailed accounts of Pepper's failures across multiple deployments, including specific financial costs and customer testimonies. There is no conflicting information regarding the robot's technical limitations.
None
Technology
Frequent Mechanical Failures
Robot prone to physical breakdowns and unplanned breaks during operations.
Inadequate Facial Recognition
Failed to recognize family members, resetting relationships after repairs.
Limited Conversational Ability
Could not carry on basic proper conversations with users as expected.
Data Inputs
Loss of Cloud Connection Data
Failed to retain user relationship data after sensor repairs.
Human Factors
Unrealistic Human Expectations
Humanoid shape led users to expect human-level intelligence.
Rapid Loss of User Interest
Novelty wore off quickly due to robot's limited repetitive repertoire.
Perceived Creepiness
Uncanny appearance and synchronized movements created dystopian feelings.
Process and Methods
Inadequate Software Testing
Robot repeatedly broke down during practice runs before live ceremonies.
Inefficient Task Execution
Robot-operated hotel failed to save money and created more human work.
Management
Overhyping Capabilities
CEO and marketing promoted unrealistic capabilities, inflating expectations.
Uneconomical Pricing Model
High purchase and subscription costs outweighed the actual utility.
Flawed Product Positioning
Positioned as a functional worker instead of a low-risk pet companion.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
