In a context of global supply chains, increasing regulatory pressure and more demanding consumers, AINIA held an online session analysing how artificial intelligence (AI) is transforming food safety. The head of digital strategic development and alliances, David Martínez, and the food safety manager at AINIA, Roberto Ortuño, agreed that AI enhances human expertise and improves decision-making. Both also highlighted the importance of having multidisciplinary teams (technology and food safety), ensuring data quality and guaranteeing the explainability of models in order to move towards a more proactive and effective approach to food control.
AI, a copilot for evidence-based decisions
For the head of digital strategic development and alliances at AINIA, AI does not replace expertise; it amplifies it. It is our copilot to react earlier, with greater precision and with evidence-based decisions. Under this premise, David Martínez explained how the combination of advanced sensing, data analytics and predictive models makes it possible to move from reactive models to proactive approaches in food safety.
In this regard, the food safety manager at AINIA highlighted efficient data management: We handle more and more data and more responsibilities. AI helps us turn data into useful information to make better and timely decisions. Roberto Ortuño added that the key lies in standardising and unifying internal and external data to make evidence-based decisions, with explainable models aligned with regulatory requirements.
Practical applications: from early warning to in-plant prediction
The speakers agreed that the most effective approach is progressive, interdisciplinary and based on rapid pilots with clear metrics. They also highlighted the importance of interoperability and sector collaboration to multiply the impact, especially in areas such as early warning and the identification of emerging risks.
Among the applications presented in the webinar, it was shown how AI accelerates the detection of emerging risks through the automated curation of signals in scientific literature, improves the filtering of early alerts (RASFF, AESAN) and enables the anticipation of contamination in plants through the integration of IoT sensors and imaging. Cases of automatic in-line inspection, data analytics for KPIs and real-time prediction of microbiological risks were also addressed.
The combination of early warning and advanced analytics allows us to anticipate emerging and re-emerging risks and act before a problem reaches the market, highlighted Roberto Ortuño.
Implementation keys: defined use cases, quality data and mixed teams
Regarding the pace of adoption, David Martínez noted: The challenge is no longer whether to apply AI, but when and with what level of ambition. Those who integrate these technologies will make the difference in food safety.
Among the operational recommendations, both speakers suggested starting with concrete and measurable cases with a clear vision of success; ensuring clean and representative data; creating mixed teams (AI, food safety and plant staff); prototyping and validating quickly with short iterations; requiring transparency, explainability and traceability of models; and promoting interoperability to avoid data islands and enable data spaces with partners and suppliers. This approach makes it possible to accelerate the transition from idea to reality and consolidate a framework for continuous improvement and more robust audits.
