[co-author: Gloria Malpass, PhD, and Heather Hatcher, Ph.D.]

Photo by Oat Canada on Unsplash

U.S. households are drinking less milk these days. Instead, alternative plant-based “milk” beverages (made from soybeans, almonds, oats, coconut, etc.) are on the rise. A study reports that around 38% of households consume plant-based drinks regularly, with 22.8% almost exclusively.

This trend is likely backed by several factors, from health needs to environmental concerns: 1) For those with lactose intolerance (not having enough lactase to digest lactose in milk), plant-based milk is welcome. About 68% of the population worldwide, and about 36% in the United States, suffer from lactose intolerance. 2) People with milk allergy — predominantly infants and young children, anywhere between 1.2% to 17% of the population — also cannot drink milk. 3) The amount of sugar and fat in milk draws some people to plant-based alternatives as a healthier option. 4) And lastly, people who are concerned about pollution and greenhouse gas emissions produced by dairy cows and their manure may look to plant-based milk as a sustainable alternative.

Despite their nutritional and sustainable features, alternative “milk” beverages have not been quite there in terms of look, taste, and texture. According to NotCo, a Chilean food-tech company, 33% of plant-based milk consumers in the US move back to dairy because of compromises in taste. NotMilk is a revolution: it looks, tastes, cooks, and blends just like milk, according to NotCo. The secret behind this product is a mixture of non-traditional ingredients decided on by NotCo’s artificial intelligence (AI) and machine learning program “Giuseppe”. The surprising ingredients include pineapple, cabbage, chicory, and coconut. NotCo rolled out its vegan whole and 2% reduced-fat NotMilk at over 450 Whole Foods across the U.S. The Giuseppe AI technology also helped NotCo create the company’s other products, including NotIceCream, NotBurger, and NotMayo.

Relying on data and networks, AI is unbiased. Combinations and proportions of ingredients that might otherwise be ignored or missed by humans are evaluated along with other potential recipes. Machine learning algorithms can analyze a large number of formulas and ingredients to help identify patterns and novel combinations that will be useful in the design of a particular flavor. AI expands possibilities for food and beverage tech companies as seen in McCormick & Company and IBM Research’s collaborative project to develop new flavor experiences using AI.

Keep in mind, though, even AI-perfected plant milk can never be dairy milk. Cow’s milk is a good dietary source of necessary vitamins, minerals, and protein. The nutritional quality of plant-based milks varies greatly, with some varieties falling nutritionally far short compared to cow’s milk. The US Food and Drug Administration (FDA) is tasked with protecting consumers from imitation products and false advertising by implementing Standards of Identity that define parameters for what we can call certain types of foods. The FDA has argued that these innovative “milk” products could mislead consumers into assuming that plant-based milk has nutritional properties similar to those of dairy milk. The FDA contends that such products should be labeled as “non-dairy beverages”, which have no presumed nutritional profile, and has invited comments (FR Doc. 2018–21200 Filed 9–27–18; extended comment period) regarding the labeling of these products with names that include dairy terms such as “milk”. If passed, the DAIRY PRIDE Act (S.792; H.R.1769), a bill introduced in the US Senate and the US House of Representatives, would prevent plant-based non-dairy products being labeled with dairy terms and would require the FDA to issue guidance in the “enforcement against misbranded milk alternatives”.

Manufacturers are advised to consult federal and state consumer protection laws regarding permissible labeling for plant milk. Labels that could be considered misleading could open the floodgates to lawsuits similar to those concerning “natural” food labels. (Juntti, 2019; Watson, 2018; Kershen, 2018; Watson, 2014)

[View source.]