Artificial intelligence is nothing new in the big food world.
mccormickThe company, which owns brands such as Frank’s RedHot, Cholula and Old Bay, has been leveraging AI in flavor development for nearly a decade and says it has reduced development timelines by an average of 20% to 25% by identifying promising flavor combinations and narrowing down ideas worth testing with physical prototypes.
It’s a similar story unileverwhere AI is deeply embedded across food R&D, with systems that can digitally test thousands of recipes in seconds and reach workable concepts with fewer physical trials. As an example, Unilever’s Knorr Fast & Flavorful Paste was developed in about half the time it normally takes. On the packaging side of the business, AI modeled how the formulation would behave inside Hellmann’s Easy-Out squeeze bottle. The company says this saved months of physical lab work.
Back in 2017, a team at Google Brain (now part of DeepMind) used AI to create the “perfect” chocolate chip cookie recipe.
But even though AI is increasingly shaping how food companies decide what ends up on grocery store shelves, they’re quick to stress that AI isn’t going to take over the kitchen.
“Human creativity and judgment will lead the way, and AI is the tool that will help us scale our impact,” said Annemarie Elbers, Head of Ecosystem, Digital and Data for Food R&D at Unilever.
“These tools help fuel the creativity of flavor scientists,” Anju Rao, McCormick’s chief scientific officer, told CNBC. Rao emphasized that AI does not replace human expertise but acts as a co-creation tool. “Our greatest asset has always been our people, who bring a global perspective, flavor expertise and human creativity,” she said.
A growing ecosystem of startups is positioning artificial intelligence as a way to approximate and predict sensory outcomes using large data sets that model how consumers will react to new foods before they are physically tested, but it’s unclear how successful their efforts will be in cracking the test kitchen code. Companies like Zucca, Journey Foods, NielsenIQ, and AKA Foods market their platforms as “virtual sense” or AI-powered systems designed to digitally screen recipes, suggest formulation changes, and predict consumer preferences before a physical prototype is created.
These companies promise much of what the food giants say they’re already doing. The idea is to create a system that can reduce the size of traditional taste panels, reduce the risk of launch failure, and shorten product development cycles by identifying promising concepts early in the process. Industry analysts predict that the global market for artificial intelligence in food and beverage will grow from approximately $10 billion in 2025 to more than $50 billion by 2030, driven by increased investment in data-driven product development, automation, and personalization.
But some early food AI pioneers are already taking the next step. McCormick’s early AI work was developed in partnership with IBM, which had previously explored AI-powered food projects such as Chef Watson. An IBM spokesperson said in a statement that the company is “no longer actively focused on this area.”
Food scientists who have tested these platforms say that behind the marketing jargon, the technology is still in its infancy, with many of the claims focused on attracting capital as well as replacing human expertise.
Brian Chau, a food scientist and founder of food science and food systems consultancy Chau Time, said many AI food startups are still in the data collection stage, working to aggregate enough real-world information to make their models meaningful predictions.
“I think every AI company that’s coming up is to some extent overstating what they can do, and that’s true of most startups,” Chau said. “Before these can actually work at scale, we need to attract investors, we need to build datasets, and we need real industry partners.”
Chau said most current platforms resemble large language models trained on existing recipes, manufacturing data and consumer trends, rather than systems that can independently generate viable new products. “When we tested one platform, the output was essentially the same as what you would get from a typical AI system,” he said. “Without proprietary data from real companies, there wasn’t much added value.”
In his view, the technology’s long-term potential depends on startups securing partnerships with large food manufacturers willing to share their internal formulation data, which many companies are reluctant to do due to intellectual property concerns. “Unless the big players in the industry input real data into these systems, it’s very difficult to make true predictions,” Chau said. “It’s a numbers game.”
Where AI food science still falls short
From a scientific perspective, the biggest hurdle is not computational power, but biology, the researchers say.
Dr. Julian Delarue, professor of sensory and consumer sciences at the University of California, Davis, said expectations for AI-driven sensory tools may be fueled by misconceptions about what AI can realistically model. “I think there’s probably some hype,” Delarue said. “That doesn’t mean AI isn’t useful. It’s just not what people expect from AI.”
While AI can help analyze chemical data and improve the efficiency of food development, Delarue said attempts to predict how people perceive complex flavors remain fundamentally limited. “When you try to predict what people will recognize from complex mixtures of compounds, the answer is no,” he says.
One of the key challenges, he explained, is that human sensory perception is inherently variable. People consume the same compounds very differently depending on genetics, culture, experience, and even personal history. “There is no such thing as an average consumer,” Delarue said. “Trying to predict what the ‘average’ person will perceive is likely a dead end.”
Lifting this restriction will require access to far more data than we currently have, at an individual level, and to know what each person or group actually knows, Delarue says. “And it’s hard work,” he added.
That variability makes it difficult for any model, human or machine, to serve as a universal proxy for taste, he said.
Even the companies building these tools emphasize that human judgment remains central.
AKA Foods founder David Sack said the company’s platform is designed to organize in-house research and development knowledge and is not meant to replace food scientists or sensory experts. “Food R&D teams accumulate a great deal of valuable knowledge, from past formulations and sensory data to tacit personal know-how,” Sack said. “But they are often fragmented and difficult to reuse systematically.”
Why humans continue to be trend makers
AKA’s platform helps teams test ideas digitally before committing to physical testing, allowing scientists to focus on the most promising formulation paths. “It’s not a replacement for food scientists or sensory experts,” he says. “Ultimately, humans define the goals, constraints, and success criteria. Sensory experts design and interpret the panels. Scientists decide what to test and what to start. AI can reduce the number of tests needed, but it does not eliminate the need for tasting and validation by actual humans. If the end consumer is a human, a human always needs to be in the loop,” he said.
“Consumers use taste to decide whether they like a product or not,” said Jason Cohen, founder and CEO of Simulacra Data, a company that uses AI to analyze sensory and consumer data. “We still start with real human sensory data. AI can only help us extrapolate insights faster and cheaper.”
Cohen, who is also the founder of Analytical Flavor Systems, which was acquired by NielsenIQ in 2025, said AI is not meant to replace human perception, but is most useful for identifying off-flavors, narrowing down formulation options and prioritizing ideas worth testing.
Chow says large food companies are uniquely positioned to benefit from AI-driven tools because they already manage vast amounts of proprietary formulation, sensory and manufacturing data. This is something that most small brands are still trying to build.
Delarue believes that the real value of AI in the food industry lies in efficiency, not creativity. This is to enable researchers to analyze data faster, manage complexity, and operate under increasing health, sustainability, and cost constraints. “Designing food today is much more difficult than it used to be,” he said. “We don’t just want to make food that people like. We need to make food that is healthy, sustainable, and affordable. AI gives us more power to handle that complexity.”
However, when it comes to taste itself, humans are still the standard. “It’s always the consumer who decides what tastes good,” he says. “It’s not a machine.”
