Podcast Ep 10: AI for Nutrition for Older Adults

Welcome to AI-Powered Caregiving, a podcast exploring how technology is reshaping the way we care for older adults. I'm Ishanya Anthapur, product manager and your host from Zemplee, where we're building AI-powered tools to support caregivers with revolutionary health tracking, alerts, and in-residence virtual care. To kick off today's conversation, we're talking about the future of food and dignity — personalized nutrition.

And we have an amazing guest, someone who's been at the forefront of culinary AI innovation. Our guest today is an award-winning chef and food innovation leader working at the intersection of food and AI. He's gained national recognition after competing on and winning a guest spot on Beat Bobby Flay — so he won against Bobby Flay — and he later helped develop IBM's groundbreaking Chef Watson AI platform. Today, he's the founder of CulinAI, where he'susing precision AI to design meals that are nutritious, personalized, and culturally meaningful. James also recently presented at the United Nations AI for Good Summit on how AI can shape the future of personalized food and health. This summer, he'll also be presenting at the prestigious Gottlieb Duttweiler Institute in Switzerland on the future of the intersection of food with AI. Welcome to the podcast, James Briscione.

James Briscione: Thank you so much for having me — excited to be here and chat with you today.

Ishanya Anthapur: We're really excited to have you. Let me start with one of the questions I'm most excited to get to. You've been in the AI and food world for a long time — since 2012, when you started to help develop IBM's Chef Watson. Could you tell us a little bit about that project?

James Briscione: Yeah, absolutely. This really began in 2012, and we're all so used to just speaking to our phones now, all of the innovations that are second nature right now. But this was coming right off the heels of IBM blowing everybody's mind by taking Watson on Jeopardy. It was showcasing natural language processing — you could ask a computer a question and it would respond — and this was something that the majority of people had never seen before or even thought was possible. And it's wild to think that was just less than 15 years ago, and it is so commonplace now.

So the next idea, after proving a computer could answer any question and conquer any Jeopardy champion, was: could the computer help people become more creative? The team decided they wanted to explore that. They thought about music and art, and then food — and food was the most tangible and the most objective route. So they came to the Institute of Culinary Education, where I was director of culinary development, and presented this idea. It started off in my mind as a man-versus-machine kind of thing. I said, let's get in the kitchen and I'll show you — I can cook circles around a computer. A computer doesn't know how to taste. It doesn't have a sauté pan. There's no way a computer could possibly know more about food and cooking than me, after almost 15 years in the kitchen.

I have a very analytical brain and a strong science background, so we talked about the creative process and how I go about creating a dish — and it was very similar to what they were trying to build with Watson. So I thought, yeah, this will be a fun little thing to do one afternoon.

Day one, I was completely blown away by the understanding that Chef Watson had about food, the patterns it could reveal, and the connections and things I was able to learn from it right in that very first session in the kitchen. It was really something — and it just evolved greatly from there.

Ishanya Anthapur: That's amazing. I love hearing these stories. It's so interesting that you also learned from it. You went in thinking that cooking and food is such a human experience — the art of it is so human — that maybe a computer couldn't have that touch. But like you said, there's something a computer can do with pattern recognition that it can do faster and better than humans. So after your experiences with Chef Watson and now with your own company CulinAI, where you're using precision AI to design meals — what do you think is the balance? Does AI have the potential to truly improve meals, or is it over-optimizing something that is inherently human?

James Briscione: (brief pause as he says goodbye to his wife)

Sorry about that. Yes — I think there are two ways to look at it. One is on the creative side as a chef. The other, and really where I spend a lot of time thinking, is on the caregiver side — how AI can help there. And I think that's really important. When you approach it not as a robot that's here to cook for you, but as a collaborator that can help streamline processes — as a caregiver, you're trying to balance so many different elements. The expectations and needs of a caregiver can include being a full-time medical professional, whether a doctor or a nurse, a nutritionist, a chef, a grocery shopper — so many different roles.

AI can balance all of that. It can look for medication interactions, flag hidden sodium issues in food being prepared, and help with planning. When you're managing so many different factors that require expertise across so many fields, being able to synthesize all of that through a single interaction can be extraordinarily powerful.

Ishanya Anthapur: That makes sense. Let me just rephrase to make sure I've got it — basically, cooking for someone is still a human act, but there's a lot to it. Like you said, with diet management, preparation, being aware of restrictions and building recipes around them — that's where AI can help. I love the sodium example. That's something I never think about, but I'd love to have less sodium in my diet. If something can recommend a recipe that's inherently lower in sodium, I'm all for it.

James Briscione: Exactly. And as a chef of 20-plus years, I know how to manage those little things — we can eliminate this, or swap this ingredient because it's going to deliver the same flavor profile with much lower sodium levels. Those smart swaps are really one of the most powerful things AI can do.

This ties back to Chef Watson and one of the first things we really dove deep into — and the reason I wrote my last book, The Flavor Matrix. It's about the chemical compounds in food that are responsible for flavor. In an apple, you've got over 400 different aromatic compounds that create its flavor. You think it tastes like an apple, but you don't realize all of the complexity that's actually happening.

Each of those compounds plays a role. Now, if you imagine that some of those compounds are tied intrinsically to fat, sugar, or sodium — and we can find another source for that compound that isn't tied to what we're trying to eliminate or watch out for in the diet, or that is tied to something beneficial like potassium, magnesium, or iron — now we can start improving the diet and delivering the same flavors.

Ishanya Anthapur: That is so interesting. The average person who's a caregiver — whether professional or personal — does not have the time or bandwidth to think about all of this when they're grocery shopping.

James Briscione: The majority of chefs I know don't necessarily have all of this either. With my background in nutrition, dietetics, organic chemistry, and biochemistry — and studying this for years beginning with Chef Watson — I have a strong understanding of it, but it's not common knowledge. There's some of this that chefs know intrinsically after years in the kitchen, but it's not being taught anywhere.

Ishanya Anthapur: So tell me a little bit about the development of CulinAI. It feels like you have this desire to transfer this knowledge and make it more accessible. Is that what drove you to start it?

James Briscione: Yes. AI is really such a great and powerful tool. One of my taglines is: AI is only as smart as we ask it to be. That's really important when thinking about your relationship with any AI system. If you know how to interact properly — ask the right questions, get the right information out of it — it can be very powerful.

With CulinAI, what we were really interested in doing was different from general chatbots like ChatGPT or Gemini, which are indexing publicly available information. We are creating a purpose-built model with a deep understanding of flavor science, recipe creation, and medical nutrition — trained specifically on those elements.

Because if you ask ChatGPT to create ten gluten-free recipes for you, one or two will probably have gluten in them somewhere. You point it out, and it says, "Ah, you're right, here are six corrected recipes" — and then it goes right back to the same mistakes. When you have a high-stakes situation like caregiving or recovery, a mistake like that in a recipe is critical. You really need a trusted source. General-purpose AI tools are not there yet. That's what we want to build with CulinAI — a very powerful culinary nutrition engine.

Ishanya Anthapur: I'm curious — how did you train the model? Did you feed it recipe books and similar publications? Given your experience, I imagine you fed it your own books as well.

James Briscione: Yes, a lot of specific data sources. The USDA publishes a lot of great information — nutrition data, recipe databases. There are a few good public recipe databases, and then we added all of my books. We also incorporated medical nutrition information — how diet-related disease works, what the typical factors are, where diagnosis lives through food. There's a lot of very dense technical data that goes along with that, plus the flavor science — large databases of ingredients and flavor compounds so that we can do that matching and those smart swaps, and find spots where we can dramatically drop the sodium in a recipe without changing the flavor.

Something we discussed a little off camera: the idea of nostalgia. The strongest sense tied to memory is smell, and smell is what flavor is. So the idea that grandma wasn't really worried about macros — she wanted meatloaf. But now we know a lot more. If we can take grandma's meatloaf recipe, get more lean proteins in it, increase the vitamins, take down the fat and sodium in really smart ways — that's an amazing thing. You can keep that comfort and nostalgia in your food, but still deliver the fuel your body needs.

Ishanya Anthapur: I'm so curious about this aspect of memory that CulinAI is working on. As someone who loves to cook and loves my mom's food, I'm very afraid of losing that particular flavor someday. The fact that CulinAI could approach preserving your memories, your recipes, and your culturally relevant meals — and transfer that to your caregiver or next of kin to cook — is beautiful. It really is.

James Briscione: Yes, and cultural relevance is such an important thing. I have a good friend who's originally from India and had gestational diabetes. She got a handout from the nutritionist with her OB-GYN listing the foods she needed — and not a single recipe appealed to her. This is not my food. This is not what I want to eat. When you talk about care, that cultural relevance in nutrition is so important. That's another big piece of what we want CulinAI to be able to do — deliver culturally relevant meals and dishes that fit the nutritional needs of anyone.

Ishanya Anthapur: I love this idea. I'm so honored to talk with you and pick your brain. The work you're doing at CulinAI is really amazing, and it's going to help improve not only the lives of caregivers but also the way people age in this country. Thank you so much, James.

James Briscione: Thank you — great to chat. I appreciate everything you guys are doing.

Ishanya Anthapur: And everyone listening — go check out CulinAI. It's an amazing platform. The website is C-U-L-I-N-A-I, and everything will be linked in the show notes, including the link to James's book, The Flavor Matrix.

James Briscione: Yes — the old-fashioned way. Open up a book. It still works!

Ishanya Anthapur: Thank you so much, James. Have a great day.

James Briscione: Thank you. Have a great day.