The ever-increasing complexity and global reach of food supply chains have amplified the challenges associated with food safety, and with that the significance of accurate, real-time food risk intelligence. A groundbreaking project is underway – EFRA (Extreme Food Risk Analytics) is spearheading an ambitious mission to revolutionize food safety measures, transitioning from reactive responses to proactive strategies utilizing predictive AI technologies.
For this ambitious goal, we need your help. We are inviting industry experts, data scientists, researchers, and regulatory bodies to join us at the 1st International Summit on Privacy-Preserving AI for Food Risk Intelligence.
Best Practices from Existing Food Intelligence Networks
Successful food intelligence networks already exist and already play a crucial role in enhancing food safety and innovation. These networks, built on strong partnerships with regulatory bodies, leverage data anonymization techniques to foster collaboration between different entities in the food sector. They are invaluable resources, helping participants to pull together collective intelligence to better understand food trends, supply chain issues, and potential food safety and authenticity problems using data aggregated from multiple sources.
In the 1st International Summit on Privacy-Preserving AI for Food Risk Intelligence, we will hear from two successful initiatives that have harnessed the power of food intelligence networks to make significant strides in food safety: FIIN and Food Fortress.
FIIN, a leading food authenticity network, has been an excellent example of successful collaborative intelligence sharing, boasting over 50 members across the EU and globally. Through their collaborative and data-driven approach, FIIN focuses on fostering authenticity and traceability in the food sector, making it less susceptible to fraud and enhancing consumer confidence.
On the other hand, Food Fortress, an industry-led initiative, has revolutionized food safety in animal feeds. It has successfully developed a cooperative approach to testing that not only increases the efficiency and coverage of monitoring but also reduces costs and risks for participants. Their model is a testament to the transformative impact of intelligence-sharing networks in fostering a proactive approach to food safety.
Transforming Food Safety Measures with Proactive Strategies
At the heart of EFRA’s mission lies a fundamental ambition—to usher in a new era of proactive food safety measures. We believe that through predictive Artificial Intelligence (AI) technologies, we can achieve a considerable reduction in food safety incidents. The potential here is not only to safeguard consumer health but also to save stakeholders considerable resources by minimizing the cost of post-incident mitigations such as recalls, surveys, and formal control activities that inevitably follow downstream food safety incidents.
However, transitioning from a reactive to a predictive approach is not without its challenges. Such a change necessitates access to vast amounts of high-quality data—data that is often privately held, sensitive, and legally complex. Food companies, understandably, are often reluctant to share this valuable information due to its sensitive nature.
To overcome this hurdle, we at EFRA are determined to foster an environment of trust and mutual benefit, enabling the sharing of this vital data while prioritizing privacy and data ownership rights. This delicate balance between information sharing and privacy protection is essential to the successful implementation of our proactive strategies.
EFRA: Pioneering Privacy-Preserving AI Training
Recognizing the sensitivity and legal complexities around privately-held data, EFRA is paving the way for a new approach to privacy-preserving-by-design AI training. We are developing a unique two-level methodology that respects data privacy and upholds the principles of trust that are essential to our work.
The cornerstone principle of this innovation is that no data leaves the company premises during AI model training. This approach involves sending an AI model trained on publicly available data to a food company’s premises. There, the AI model is further trained on the local company’s data, leveraging their computational resources. The result is a personalized AI model that can predict company-specific ingredient, hazard, or supplier emerging risks, without exposing sensitive data.
The second level involves the trained AI model leaving the premises and moving to other food companies to be trained similarly. It allows the model to encounter diverse examples from multiple organizations, enhancing its predictive power while ensuring that data privacy is maintained. Only the AI model moves around, eliminating the need for direct data sharing. This final model is then distributed to all participating companies, further bolstering their predictive capabilities.
Through these innovative measures, EFRA is championing a new approach to proactive food safety—one that maximizes the potential of predictive AI technologies while preserving the sanctity of data privacy. We firmly believe that this strategy will lay a foundation for the future of food risk intelligence, ensuring a safer and healthier food supply chain for all.
We need your help
To realize these ambitious goals, we need active collaboration across a spectrum of stakeholders. The EFRA Platform will showcase the real-life applicability of our technologies through use-cases targeting poultry pathogen prediction, risk prevention, and food-safety-optimal pesticide use. This is why we are inviting collaboration at the 1st International Summit on Privacy-Preserving AI for Food Risk Intelligence.
As we work towards the creation of a more secure and proactive food industry, we invite you to bring your expertise and join us in shaping the future of food safety. Let’s collaborate to enhance the development of trustworthy, accurate, and fair AI systems for food risk prevention. Together, we can create a safer, healthier future for food consumption worldwide.
If you are interested to learn more about relevant initiatives and contribute your own expertise in interactive brainstorming sessions, please consider registering in the link below.
