From boosting their mood to powering performance, consumers want more from their foods. But for ingredient suppliers, meeting those demands isn't as simple as it seems.
It can take decades to find ingredients that are not only clinically proven to provide health benefits, but also will remain stable when added to food or beverages. To speed up the discovery process and keep up with consumer wellness demands, more companies are leaning on technologies, such as artificial intelligence.
Biotech firm Nuritas is among the pioneers in AI ingredient discovery, using the technology to identify natural peptides that could have functional benefits, including in food. Since its founding in 2014, the company has developed ingredients to build muscle, improve sleep and support youthful-looking skin. Nuritas has even caught the attention of big food companies, leading to partnerships with companies including Nestlé and Mars.
Nora Khaldi, founder and CEO of Nuritas, spoke with Food Dive at the Future Food Tech conference in San Francisco about how AI is transforming food science and shaping consumer demand for functional ingredients. Khaldi said AI has allowed the food industry to understand ingredients at the molecular level, which could pave the way for pharmaceutical-like ingredients in food.
This conversation has been edited for clarity and brevity.
FOOD DIVE: How is AI shaping ingredient discovery? How does Nuritas use the technology?
KHALDI: What used to take decades to develop is now taking months. So we can do it from start to end, from starting with a concept to having something in a human, you're talking a few months, not counting the clinical trial.
We believe that everything is found in nature. It's just that as humans, we haven't been capable of data mining it the right way and accessing it the right way. So that's what technology has allowed us to do. We're going into humans with a very high confidence.
Our muscle building ingredient PeptiStrong has five major peptides that would have taken 30 million years to discover with the current techniques of time versus the AI. Not only that, but our success rate in humans so far has been about 80%. That's unheard of, and it's because of AI shaping the downstream elimination process, filtering for peptides that are stable, orally available, will resist digestion and resist heat if the food will be baked.
We had to fundamentally spend eight years on R&D. AI is only used on data, so you have to still build that data. But after we built it, we realized the AI now is capable of being smarter than our scientists and can understand the biological needs of the human body and connect them to a molecule in nature, a peptide.
One of the big worries among consumers around AI is that the technology can hallucinate or get things wrong. Is that a concern for you?
AI has zero value without validation. Just like in the pharmaceutical industry, validation is everything.
What AI simply does is it allows us to see more in the natural ingredient world, more things we couldn't see as humans. You can't just jump from a prediction and say, "Okay, now I'm going to put this into food." You have to say, "Okay. This molecule is found in rice husk or fava beans and when broken up there are peptides that actually heal the body. How can we access and give those accessible materials to humans?"
There's a lot of functionality within food that our bodies cannot access. So AI allows us to just find those and then there's a battery of validation including clinical testing and regulatory approval. That battery of tests is not going to disappear. So it's all about validation. It's just AI is helping us to find it.
You've mentioned that AI is helping the food industry better understand ingredients down to the molecular level. Why does that help with the formulation of functional foods?
The big question for everyone is, ‘How do you take processed food and make it more nutritionally relevant for the individual, so that we're eating something that's still tasty, but it's not causing harm to us and could even help us?’ I think that's the principle of what we're trying to do.
Originally, the food industry has always been about macronutrients. So they do clinical trials with macro, and you get guidance like "eat more fat," or "eat less fat." It's also why you see things saying coffee is good for you one day and it's bad the next day. These conflicting things happen because macro level doesn't give you the right picture.
It's the molecular level that gives you the right picture. Through AI, we can now understand the molecular level of food, so we can better design the clinicals. We have more definitions of what we're actually testing. You understand why things are working or why they don't work, why they work in a certain population, why they don't work in others. So it allows us to see better, to prove things and avoid conflicting results.
Food is so complex, there are so many billions and trillions of molecules in everything you eat. It's not a pure molecule. But now with AI, you can literally go down to that level and understand a little bit more. We understand what we're actually testing with the molecules in that so we're not coming up with a general result of "these fava bean ingredients are good for you." Instead, we can say "these peptides are really good for strength."
How is AI shaping functional ingredients outside of the lab? Could consumer AI technology shape demand for certain ingredients?
It's going to be transformative. And honestly, CPG companies don't know how transformative and how fast this is coming at them.
You can input all your data from your doctor and blood tests, and AI literally tells you what ingredients to avoid what things you should be eating already today. With smart glasses, you can look at a product and it will tell you "these ingredients are toxic," or "avoid this."
Wearables like smart watches and rings can track sleep and cortisol and blood sugar spikes. It's allowing people to make better choices, like now people can actually choose better foods based on what their body needs using their wearable data.
So it's no longer the brand that counts, and what the storytelling of the brand is like. Because a consumer could think it's very healthy and then immediately see a massive spike on their Apple Watch, allowing them to contest those brand's claims in real time.
And I think that really speaks to a larger point that is: Functional is no longer throwing in a few vitamins. That's past. Now it's really about clinical proof, scientific proof, things that are totally unique and measured.
