Identifying environmental factors associated with tetrodotoxin contamination in bivalve mollusks using eXplainable AI

Explainable AI can help protect food safety: Researchers used an interpretable deep learning model to predict tetrodotoxin (TTX) contamination in shellfish, demonstrating that AI systems designed to show their reasoning—not just make predictions—can be trusted tools for real-world food safety decisions by industry and regulators.

Sunlight exposure is a key contamination driver: The study identified day length and global radiation (sun hours) as among the most important factors associated with TTX contamination in bivalve mollusks, suggesting that seasonal and weather patterns play a critical role in when shellfish become unsafe to eat.

Multiple environmental factors work together: Beyond sunlight, water temperature and chloride concentration also significantly contributed to contamination risk, indicating that TTX contamination results from complex interactions between multiple environmental conditions rather than a single cause.

Early warning systems are now possible: By identifying which specific environmental factors predict TTX contamination, this research enables the development of practical early warning tools that authorities and food producers can use to prevent contamination before it reaches consumers, reducing both health risks and economic losses.

Transparency in AI builds confidence in predictions: The researchers specifically chose an "explainable" AI approach rather than a black-box model, allowing stakeholders to understand why the system makes its predictions—a crucial requirement for responsible AI use in high-stakes decisions affecting public health.