This research introduces RECALL-MM, a novel multimodal dataset built from consumer product recalls that enables data-driven risk assessment in engineering design.

• Transforms historical product recall data from the US Consumer Product Safety Commission into actionable safety insights
• Augments traditional recall data with generative methods to create a more robust analysis framework
• Identifies recurring patterns in product failures to highlight specific areas needing improved safety measures
• Demonstrates how computational methods and LLMs can predict and prevent potential hazards

For engineering teams, this approach offers a powerful way to anticipate safety issues early in the design process, potentially reducing costly recalls and enhancing consumer protection.

RECALL-MM: A Multimodal Dataset of Consumer Product Recalls for Risk Analysis using Computational Methods and Large Language Models