Can AI help us build a more sustainable food system?

Argentina and Brazil have emerged as important producers of both raw and industrialised foods, including meat, soybean, sugar cane and corn, among other staple products. But their meteoric rise as food producers came at a cost: agriculture is now one of the main drivers of greenhouse gas (GHG) emissions in Latin America, with nearly 40 percent of gases such as carbon dioxide and methane, which contribute to climate change, emitted in the region originating in agriculture, forestry and other land use. 

Considering this data and the prospect of further development of the food sector within the next few years, the Food and Agriculture Organization of the United Nations (FAO) states that investing in a green transition is essential in order to grow sustainably.  

Going vegan has emerged as the ideal way to lower our food-related carbon footprint. Studies carried out by scientists from Stanford University and the University of California point out that, in the most ambitious scenario of meat-based diets transition, up to "90 percent of the emission reductions could be achieved by only replacing ruminants such as cattle and sheep." 

But despite the undeniable benefits of adopting a vegan diet, animal product consumption is expected to rise over the next decade, FAO predicts. Notwithstanding, vegan and plant-based industries foresee a bright future - with a multi-billion dollar market under development. 

Lilia Calheiros de Oliveira Barretto, a food engineer and professor at the Federal University of Sergipe (Brazil), investigated the ways in which artificial intelligence can be used for plant-based food processing, since, according to her assessment, it is the future of the food industry. She said that several companies have been investing over the past decade in the development of new technologies to produce better vegetable products, recognizing the immense potential of this industry.

Start-ups and food companies in general have been betting on substitutes such as chickpeas or other vegetables to replace animal protein in ice cream, pudding, mayonnaise and even the meat itself 

The nutritional value of these often highly-processed, sodium-rich alternatives, however, is the subject of ongoing debate.

"Today there are many industrialised products based on vegetable protein with reduced sodium content as well,” the researcher pointed out. "These innovative industries that have been incorporating vegetable protein in their composition have also been concerned with this composition issue in terms of abusive sodium content."

She added, "We have [in Brazil] a new legislation that came into force from last year that already provides for changes in the labelling of processed foods [to inform about excess sugars, sodium and fats, for example, following the example of other countries]. So what the industries are doing today is to seek changes in formulations to try to avoid the presentation of this new visual identity required by legislation."

Nevertheless, these products contain a decent quantity of protein and offer a flavour eerily similar to that of meat,  which can help ease the transition for those intent on weaning off of animal products.

According to a study published by Veganuary, an initiative aimed at encouraging people to stop eating animal proteins, Chile saw the highest increase in the supply of products labelled as vegan recently, which now represent 12 percent of all food products  in the Latin American nation. 

Chile is also home for a start-up that has been transforming the way some brands think about vegan food by exploring plants’ ability to create enjoyable flavours of milk, meat, mayonnaise and other day-to-day products. Now a B certified company for its commitment to reducing GHG emissions and water use, the company, which was founded in 2015, works in collaboration with other internationally-established food companies to offer vegan in their respective markets. 

Operating under the slogan why not?, the company aims to explore "the infinite world of plants." It does so by experimenting with more than 300,000 additional plant species and varieties hitherto unexplored by humans.

Combining such a big amount of data would take a professional a long time, but AI can do the job in a matter of hours with significantly less financial input. The technology developed by the company aggregates in various vegetable and animal matrices and their sensory characteristics, as well as chemical components, Barreto explained. 

From a database, the software will then present the best formulations within its portfolio to achieve a product with a taste similar to that of an animal product. 

The future of food is tech?

The Brazilian professor believes that in the near future other companies will be able to employ similar technologies, which are already available in the market, to create plant-based foods with taste and texture resembling meat, milk and other animal products. AI can also revolutionise other aspects of food production, from farms to factories, the researcher added.

On farms, artificial intelligence can help forecast rainfall or heat waves, as well as identify potential problems in crops that need extra attention, and thus avert losses and boost productivity. Data will be consolidated in softwares, which would facilitate access. 

Big data can also be useful for speeding up the production process, reducing contamination, improving precision in the selection of raw materials and their processing, enhancing nutritional performance, and creating more specific products for each niche market. 

Tracking food origins

AI may also be used to produce a tracking function designed to assist consumers. At Brazil’s State University of Campinas, a group of researchers developed a tool based on mass spectrometry and machine learning techniques capable of detecting pesticides in fruits and vegetables. First trials of the technology with tomatoes identified chemicals in products sold as "organic," that is, free of non-natural substances. 

The next step of the project will be to make the technology applicable to all sorts of fruits and vegetables, as well as reachable by producers and consumers.

According to one of the project’s lead researchers, professor Rodrigo Ramos Catharino, AI may also help improve agricultural production through the use of drones that would prevent the depletion and waste of food resources. In the future, he believes, this set of new technologies will be readily available to all food producers and farmers, and generate cheaper and more nutritious products for the population. 

In a country poisoned by pesticides, promoting the widespread use of real organic foods is a matter of public health.

Another food-related beneficial use of AI is a cell phone-based feature that tracks meat and assesses whether a certain beef being purchased at the supermarket is involved in deforestation, forest fires, work conditions analogous to slavery or animal well-being fines. The initiative, called Do pasto ao prato (from pasture to the plate, in free translation), is supported by NGOs such as Greenpeace and financed by Norway's International Climate and Forest Initiative. It can help meat-eaters have a more transparent relationship with the animal proteins they consume and connect urbanites to the sources of their food.  

Humans and machines working together

But despite the evident breakthroughs accompanying its inexorable stride, AI should be deployed with caution, specialists point out - certainly when it comes to food. "[Negative consequences may include] choice of food, products and diets without direct consumer will, which can generate dependence without critical assessment," Catharino warned. 

The use of AI in the food industry is not designed to replace research, development and innovation laboratories run by humans. Nowadays, the role of professionals is still crucial in order to feed the database accurate, reliable information and test new recipes generated by AI in an industrial experimental kitchen: people with trained senses of smell and taste are essential to the proper development of new products and to ensure quality control. 

However, professionals will need to adapt and relearn abilities in order to thrive in this new environment. "We need to extract the best from artificial intelligence. The human being will need to restructure," said Barreto. 

The researcher further highlighted that, despite AI being a present reality, future development will require new investments in connectivity, as well as a significant drop in machine prices. "We reinforce here that the world population will grow by 2 billion people by 2050, according to estimates from the United Nations, leading to a necessary 60 percent increase in food productivity. In this sense, some challenges need to be faced so that the transition from our current situation to the big data era takes place efficiently."

Image by ThisisEngineering RAEng.