Taco Bell's AI voice ordering system at drive-throughs was exploited by pranksters who ordered 18,000 water cups and caused other malfunctions, forcing the company to reconsider its AI deployment strategy.
Taco Bell deployed voice AI technology at over 500 drive-through locations across the US since 2023, with the aim of reducing mistakes and speeding up orders. The system successfully processed over two million orders without major issues. However, viral pranks exposed significant flaws in the AI system. In one incident, a customer ordered 18,000 water cups, which appeared to crash the system. In another viral video viewed over 21.5 million times, the AI repeatedly asked 'and what will you drink with that?' when a customer ordered a large Mountain Dew. These incidents sparked widespread online ridicule and forced human workers to intervene. Chief Digital Officer Dane Matthews acknowledged that the AI 'lets me down sometimes' and announced that the company is now considering hybrid stations where human oversight can monitor and intervene when necessary. The company is developing coaching protocols to help teams determine when to use voice AI versus when to have humans take over, particularly during busy periods. Despite the viral failures, Taco Bell reports that the system has successfully processed two million orders since its introduction.
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.
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
No population impact data reported.