The hypothesis of the technical singularity proposes artificial intelligence will trigger runaway technological growth. This so-called “intelligence explosion” will result in unfathomable changes to human civilisation.
“The singularity will happen through food because we all eat every single day,” Matthew Lange, a lecturer at the University of California, predicted. “Food touches every aspect of our lives. Cooks are the original hackers and recipes are the original source code. I think we are going to go full circle with that.”
The Internet of Food
Lange is working on the development of “the internet of food” by creating a common language that can be used to digitise every aspect of the food system. Lange equates this to the development of Hypertext Markup Language (HTML) that is used to create websites. “We are building the Hyper Food Markup Language,” he explained.
“We are building ontologies here… We are doing it out in the open to democratise it. We are building lots of different kinds of ontologies about food… We are building this knowledge base across the whole value chain.”
This extends from the relatively simplistic task of categorising and naming the components of food to the much more challenging objective of developing “computable models of flavour”.
“Thinking of computable flavour is really hard… We have digitised light and sound and now we are building that language to describe flavour. But flavour is a wicked problem because it is subjective,” Lange said at the Seeds and Chips food tech conference. In particular, he noted that there are a number of variables that will alter how a person perceives flavour, from mood and personal history to expectations and environment.
AI development supported by investment
There has been an influx of investment into companies developing AI solutions.
Data from CB Insights shows $4.871bn was invested in AI companies in 2016 - a 698% increase from 2012 levels.
Digitalising flavour for personalisation
This type of tool can be used to meet some of the largest challenges currently facing the food sector, including improving population health, Lange suggested. For example, combining this computerised understanding of food with genetic data would enable a deeper understanding of health and preference. “We should not be doing the greatest good for the greatest number [through dietary advice] anymore. We have the tools, we are building the tools, to do much more.”
Johan Langenbick, co-founder of AI developer Foodpairing, also stressed the potential that artificial intelligence has to deliver improved nutrition. “AI is creating better food systems and improving nutrition,” he argued.
“On the one hand, the food value chain is out of balance. Climate change and the health challenge are on the other. Diet is at the centre of this.”
Foodpairing is focused on using artificial intelligence to understand food preference. This is key to changing dietary habits, Langenbick suggested. “If you want people to eat different, people won’t do it unless it tastes [good]. We have been digitalising flavour for ten years… We are creating a recommendation system using flavour technology to deliver personalised food recommendations.”
The group is developing a platform that takes data like what is in the fridge and, using a personalised flavour identity, “generates a novel recipe” tailored to individual preferences and dietary requirements. “We are at the intersection of culinary flavour, data science and food tech,” Langenbick explained.
The company has the “largest flavour database in the world” that includes, for instance, 8,000 aroma molecules. Foodpairing’s algorithms can identify food molecules that are similar in structure “that you wouldn’t usually combine”, like strawberry and asparagus. But the real “breakthrough” is taking this data and translating it to reflect personal taste preference, Langenbick suggested.
“The machine doesn’t know how the person receives this. Our breakthrough is to go from machine input to understand how a human perceives it… We have made a major breakthrough there.”
Alessandro La Volpe, president of IBM cloud Italy, said this kind of technology will transform food sector innovation. “Companies will be forced to radically optimise innovation,” he suggested.
IBM’s own AI programme, Watson, has been developed to dissect flavour in a similar fashion. In a recent collaboration with Virgin Holidays, Watson analysed social media data to determine what flavours consumers associate with ‘holiday feelings’. This was then analysed and translated into Virgin’s bespoke ‘holiday rum’.
Kishan Vasani, co-founder and CEO of DishQ, is also working on flavour analytics tech that can be used to inform innovation. The company has developed a “food brain” and DishQ uses this tech to provide personalised recommendations for restaurant menus.
The next stage in DishQ’s business model is to extend the application of its food brain to the FMCG space. Vasani told FoodNavigator that the group will launch a beta version of this technology within months.
“Something coming up in the near future is, can we use this food brain to tell… FMCG brands what the next thing they could launch is?”
Efficiency and environmental impact
Artificial intelligence can also be used to boost efficiency and reduce the environmental impact of food production.
This tech is currently being developed to deliver on-farm improvements through smart farming solutions.
“Most of the work that has been done so far is on optimising output,” Zachary Fritze, CEO of Promethean Labs noted.
Promethean develops AI solutions that improve agricultural efficiency. “We use AI to model threats. We combine [information from] satellites and ground level data,” Fritze explained. “We are here because of the global challenges we will all have to face… AI is a solution where we can get faster, cheaper solutions that help everyone.”
This development helps unlock the door to precision spraying that minimises the use of pesticides and helps build soil health by understanding diseases and threats. It also has significant ramifications for the farming insurance sector, which will be able to align products more closely with the needs of individual farmers.
Looking to future ag tech developments, Fritze predicted that we will see the development of “the Amazon of ag tech” – a marketplace that allows farmers to obtain objective information. He also expects to see a growing number of AI robotics companies that utilise edge computing to deliver “faster and more actionable outcomes”.
“These developments are going to create farming as service companies,” he predicted.