Food safety is a critical aspect of food product development and production within the global food and beverage (F&B) industry. The journey begins at the first process known as primary production, then proper handling, and proper transportation to distribution centers and finally on to retail supermarkets and food service tables. Food safety is, therefore, a critical aspect of the so-called supply chain management. In recent years, food safety has been compromised in many instances that have resulted in food recalls and foodborne illnesses. Consequently, consumers and regulatory authorities are continuously requiring more insight into the origin of food. Whilst manufacturers must meet consumer demands, it is of paramount importance that safety and quality of their products is ensured. Clearly, manufacturers must ensure that all products meet industry, regulatory and consumer specifications, both locally and internationally. Subsequently, the amount of data to gather and to track down grows exponentially.
Questions that arise include: What is the solution for the maintenance of food safety across the entire production chain while adhering to approved food safety standards, laws and regulations? Recently, there has been a growing interest in what is known as Artificial Intelligence (AI) and Internet of Things (IoT). Could these innovative concepts actually provide F&B companies with new supply chain insights to stay ahead of the curve and maintain food safety?
What is Artificial Intelligence?
The term artificial intelligence, is not at all new; it is known since the 1956. Nowadays referred to as the intelligent machine, AI is in fact an umbrella term for 3 distinct capabilities which include machine learning, natural processing and optimization. AI has become more popular due to the increased volumes of data, advanced algorithms and improvements in computing power and storage, that are applicable in various types of industries. Interestingly, it has even been considered as more powerful than the human brain by AI enthusiasts.
In lay terms, AI is based on programming computers so that they are able to
In fact, such technology facilitates the F&B industry to improve food safety preventive control measures and quality assurance, and it is being implemented in every sector of the industry – from perishable foods (e.g. dairy products) to dry products (e.g. grains). Its advantages and drawbacks, which are further described, have been nonetheless debatable.
Benefits of AI to the food industry
In a first scenario, AI could ensure food safety within the F&B industry due to its sterile properties, which could prevent manual food cross contamination, handling and packaging by human workers [who could be sick (unreported)]. This facility’s AI advanced ability to reduce/control/manage/prevent potential food recalls and foodborne illnesses resulting from microbiological, chemical and physical contamination would be a significant improvement. These automation machines are also convenient to clean. Technavio predicted that the application of automation/robots with the food industry will rise by 2019. In that regards, electric noses or tongues (electronic chemical sensors), which actually can mimic the human scents, can be useful to, for example, recognise odors in raw materials, and thereby improve quality control during processing. Furthermore, in this new era of innovation, according to Infiniti Research, AI could help grow better food, mainly from an agricultural perspective. At the farming level, AI could help detect plant disease, pests, and improve soil health.
Limitations to AI to the food industry
Despite the benefits of this state-of-art technology, is AI really the next big thing? Surely, it poses certain limitations that need to be considered and are described below. Firstly, it is increasingly grabbing the attention of researchers that AI is susceptible to attacks. Just like any other computer software, AI is also vulnerable to what is called adversarial attacks. They are intentionally designed to cause the machine to deliberately make a mistake. It is about misleading AI algorithms to give a wrong answer. In this case, can the AI machine be trusted?
Secondly, AI requires large amount of data to perform effectively. There are two existing AI models for task-specific training data, including ImageNet and CIFAR-10 image databases. They compose of 1.2 million and 60 thousands data points, respectively. Imagine how tedious, detail-oriented, slow and expensive this technology would be! Hence, a question that can arise is: how safe, effective and efficient are the storage, access and retrieval of such huge amount of data on a daily basis? This factor could be potentially detrimental to a food processor which manufactures various product lines. Supervision will be required to strictly monitor the accuracy of each and every single data bit being entered into the system. There should not be any margin for error in this case, as there could be serious risk of contamination or mislabelling if any single incorrect information has been entered into the automated system. Consequently, food safety is compromised, thereby causing severe damage to brand reputation and consumers’ health.
Eventually, as every machine carries a probability to fail, AI and automation can as well. This is because it can only perform the tasks that it is actually programmed for. It basically cannot make judgements. Therefore, from a food safety perspective, whatever ingredients are recorded in the system will be accepted by the machine and therefore, could not trace back if there is an allergen in that particular food product lot. Consequently, AI fails and could potentially cause a food safety hazard. In the aspect of failure, the recent car accidents that were under autopilot by Tesla, Waymo and Uber, demonstrates the hypotheticality of AI technology.
What is the role of AI and IoT on food safety issues?
Technological and scientific developments in the last decades have granted the food industry with new tools to better control and improve food safety. With these new features, the industry addressed some problems that can be traced back to the industrial revolution, where the increase in the production of large-scale consumer goods, in larger manufacturing units needed to be transported to consumers increasingly far away from production sites, posed challenges never faced before.
The question is, are we winning the battle of providing safe food to consumers? The answer to this question divides experts’ opinions and can be the subject of a whole new article. Developments in the use of AI and IoT to tackle food poisoning and improve food safety will certainly convince more people to join the YES team.
Have you forgotten to wear masks?
In some Shanghai restaurants this will not happen. The Shanghai Municipal Health Agency signed a contract with KanKan to install in 200 restaurants (with the possibility of expanding to 2.000) cameras and software that utilizes AI to instantly analyze the images obtained and detect if workers are wearing hats and masks.
Cameras and AI are also being tested and used in agriculture. One example is the use of robots with cameras that can detect weeds and spray herbicide only when and where necessary. According to ecoRobotix, the Swiss company developing this solution, the robot will reduce by 20 times the amount of herbicides farmers use. With this approach and similar ones, the amount of chemicals used in agriculture would be drastically reduced as the hazard to the food industry, not only by product direct exposure but also by soil and water contamination.
AI can also contribute to improve cleaning and sanitizing. The called Self-Optimising-Clean IN-Place developed by University of Nottingham researchers uses ultrasonic sensing and optical fluorescence imaging to measure food residue, as well as microbial debris within tanks, vessels or pipes. Data will be fed into an intelligent decision-making support tool that will assess the extent of fouling and stop the detergent phase when the equipment has been cleaned to the required level. Besides the productivity gain (cleaning time can be reduced by 50%), real-time assessment of cleaning requirements and reducing the use of water, energy and chemicals are other benefits of this system.
Hand washing is another daily and frequent GMP Good Manufacturing Practices – Personnel Hygiene Practices activity. This activity’s efficacy assessment is critical. PathSpot developed HandScaner to scan employees’ hands to detect indicators of pathogens that cause foodborne illness in under 2 seconds. Using fluorescent light spectroscopy as a way to detect the presence of contaminants this device gives real-time feedback on the effectiveness of hand washing and a dashboard to view both the frequency and the efficacy of your business’s hand washing practices so you can ensure a positive sanitation culture.
Will top management be the auditor’s eyes in an organization?
Audits and auditors will also benefit from new technology developments. In the future audits will not be performed as traditionally. One of the core principles of any food safety system is Top Management commitment. One of the most interesting developments that the future can bring is remote audits (in ISO 19011:2015 – Guidelines for auditing management systems this possibility is already addressed). When we look at the work of EyeSucceed we can foresee that audits done remotely can be more that just doing the opening/closure meeting or documentation analysis. Using smart glasses technology, audits can be performed “through livestreaming, two-way audio/visual communication that allows expertise located anywhere in the world to be brought live inside the audit site. Can you imagine how would this impact in top management commitment if internal audits were done by them wearing a smart glass and being guided through the facilities by the auditor and viewing what he is seeing? After the audit, the recording could be used as evidence or even for training purposes”.
Internet of things is the way to achieve my dream
In my recent article “Blockchain in food safety, are you ready?”, I shared one of my dreams. I thoroughly believe that we can and should reach a point where expiration date can be calculated according to the farming conditions of its raw materials, the specific conditions and operations done by the industry and how it was held during transport and retail. Most importantly, this information will be available to in a screen on a smart refrigerator door that would alert me as soon as any of my food is under a market recall. With the development of the use of IoT into the food industry, we will be able to integrate information throughout the supply chain using different technologies such as smart temperature sensors, pathogens identificators and other tools as presented above. If all this information can be associated with the product and retained in a secure cloud base network such as Blockchain, we can only imagine the insights that the F&B industry would gain about their raw materials and the increased confidence consumers would gain when purchasing/consuming food products.
This article was written in collaboration with Vasambal (Sam) Manikkam
Disclaimer: The information contained on this article is based on research done in the last months and the authors personal experience and opinion. It is not intended to represent the view of any organization they work for or collaborate with. The authors will not be held liable for the use or misuse of the information provided in the article.