AI seems tailor made for food safety. Both the food supply chain and the catering sector generate copious amounts of data that is dense with meaning. But interpreting it needs a lot of leg work.
Now we have a technology that can both capture that data and glean intelligence from it, effortlessly – like a seasoned EHO casting their eyes around a storeroom. So, is AI a fantastic opportunity for public protection? Or is it a threat? Is it going to put a lot of EHOs and food technicians out of businesss?
One would imagine that an AFSO (automated food safety operative), could carry out an inspection in nanoseconds. It wouldn’t need to set foot on the premises. In fact, there wouldn’t be an inspection.
Smart fridges, ovens, extractors and chillers, connected by the internet of things, multiple sensors and object recognition-enabled cameras able to detect grease, bacteria, insects and rodent movements and droppings, would provide a stream of real-time data back to HQ.
A little red light would blink and a text would go to the service head at the slightest sign of an issue. A well-designed algorithm would predict a problem brewing before it happened. And it would factor in customer ratings, complaints and social media posts.
We’re probably closer than we might think to the AFSO, says Helen Statham, head of health, food and fire safety at SSP Group, especially in food manufacturing.
Statham was formerly risk and compliance lead for a high street fast-food brand and served as director of food safety and trading standards for a London borough. She says, comparing the robot and the human: “In my opinion, there’s space for both. But EHOs do need to think about their role differently and envisage a future in which AI is involved.”
Importance of behaviour
Intelligence, she notes is not necessarily ‘intelligence’. Compliance is important, but the largest determinant of food safety is behavioural – that’s why one shift in a fast-food outlet can be exemplary and the next chaotic. CCTV is already widely used by food business, for security purposes and to generate evidence to defend legal claims.
It’s not much of a jump from there to using it for food safety – how often do staff wash their hands? Are they putting their blue gloves on? How often are tables cleaned? What do they do at the end of a shift?
Here, we get into a potentially tricky area of compromising staff privacy. But Statham says: “The intention shouldn’t be to tell them off and it’s not an alternative to an audit or an EHO visit. Too often, as EHOs, we’re telling people where things go wrong, but we’re not understanding why it went wrong – the behaviour that led to it. That’s where technology comes in – as a tool for improvement.”
For large brands, she notes, the risk of reputational damage is a powerful motivator for compliance. For small businesses it often isn’t. For them, Statham argues, there is little evidence that enforcement and routine inspection work – fines for food safety offences are so small that they are regarded as an occupational hazard.
Food safety training is not mandatory for those setting up a business and they know that they won’t be inspected before they open. These issues need to addressed, regardless of technology.
In Statham’s ideal scenario, tech-enabled enforcement could free up time for highly-skilled EHOs to train the ignorant and pursue the ill-intentioned – the ‘bad actors’. There will always be enough work of this nature to keep them busy.
Data into action
Food safety expert, Sterling Crew, president of the Institute of Food Science and Technology, agrees. In his view, AI won’t supplant human food safety professionals, but enhance what they do. He says: “What AI does is to translate data into information, provide analysis and guide corrective action. it’s a great tool for EHOs to use, but it’s not a replacement. The need for human intervention won’t go away. But that invention will be smarter.”
Crew is excited rather than challenged by the future. We are still only scratching the surface, he says, of what is possible. Emerging technologies and sophisticated affordable equipment, such as hand-held mass spectrometers, once only available in professional labs, are revolutionising food safety, in the catering and manufacturing sectors. They are also great for eliminating fraud from supply chains and for reducing waste. Science fiction is becoming science fact.
He says: “I have been in the food sector for more than 40 years, starting my working life as a regulator. I expect to see more changes driven by AI in the next decade than at any time in my life.” He concedes that professional bodies, technologists and scientists are often viewing AI with concern, fearing for their jobs. But, with resources in food safety enforcement so tightly stretched, he argues that they have more to gain than to lose.
Technology evolves, humans adapt. Could it be that the main problem with the AI revolution is that it’s happening so quickly? This article was written by a human, by the way. Or was it? Crew says: “We are at the beginning of an AI revolution in food safety. It has the potential to have the same impact as the industrial revolution, when blue-collar workers’ lives changed forever. AI will have a similar impact on white-collar workers. Including regulators and food technologists.”
This Blog was written for RHE Global
Will Hatchett has been a journalist since 1986 He was editor of Environmental Health News from 1998 until 2018. The views expressed here are purely his own