ChatGPT Can Plan Your Low FODMAP Meals. But There's a Problem…
[Originally published December 2024. Updated August 2026]
If you're following a low FODMAP diet, you've probably already discovered that there is no shortage of information online. There are recipes, food lists, apps, Facebook groups, TikTok videos, YouTube and now, increasingly, AI. Ask ChatGPT to create a low FODMAP meal plan and within seconds you can have a week's worth of breakfasts, lunches and dinners. Sounds pretty handy, right? But there's a catch.
Back in December 2024 (read the original blog here) we decided to put AI to the test. We asked ChatGPT and Microsoft Copilot to create a five-day low FODMAP meal plan, complete with recipes and serving sizes. The results were interesting. Some of the suggestions were genuinely useful. AI came up with new meal ideas, made sensible ingredient substitutions and understood some of the basics of low FODMAP cooking.
But it also made some confident mistakes. And that's the problem with using AI for the low FODMAP diet. As we learned in our original blog, AI is brilliant at ideas. It can help you:
think of new meal ideas
use ingredients you already have
adapt a recipe to suit your preferences
create a shopping list
suggest substitutions
plan meals for the week
use up leftovers
turn a handful of ingredients into dinner
Our 2024 experiment showed that AI was particularly good at producing simple recipes with relatively few ingredients. It also correctly picked up some obvious low FODMAP swaps, such as lactose-free milk and yoghurt, garlic-infused oil instead of fresh garlic, and gluten-free or sourdough bread. But… it’s not helpful in determining whether something is low fodmap or not. And there's a very good reason why.
AI doesn't have a FODMAP laboratory in its pocket. This is the bit that is easy to overlook. Generative AI doesn't independently test foods for FODMAPs. It generates answers based on information it has learned from existing sources. That information can be incomplete, inconsistent, second-hand or out of date. It can’t directly access the proprietary food database from the Monash Fodmap App or report on the cut-offs of specific serving sizes. Plus, one of the biggest mistakes we saw in our original AI experiment was that the chatbots sometimes appeared to treat FODMAPs as though foods belonged neatly in one of two categories:
Low FODMAP ✅ High FODMAP ❌
But that's not how the low FODMAP diet works. Serving size matters. A food can be low FODMAP at one serving size and become high in FODMAPs at a larger serving.
This is why the Monash App provides specific quantities rather than simply telling you whether a food is "good/safe" or "bad/unsafe". Monash explains that some foods are naturally low FODMAP in a typical serving but become high FODMAP when eaten in larger amounts, while others can be low FODMAP when eaten in smaller portions.
And this was exactly where our AI experiment fell down. Some of the recipes looked perfectly reasonable until we checked the quantities. We found excessive amounts of ingredients including red capsicum, broccoli, zucchini, cherry tomatoes, ripe banana and asparagus. AI had got the general idea right but missed the important detail: how much was actually in each serving? For someone experienced with the low FODMAP diet, these mistakes might be obvious. For someone newly diagnosed with IBS, they may not be.
The low FODMAP diet is constantly evolving. Monash regularly tests and retests foods because FODMAP composition can change and testing methods have improved over time. Just last week, updates to the App changed serving-size and traffic-light information for a number of foods such as vanilla essence, vanilla extract, algae, mustard powder and Vietnamese mint! And this is important. Imagine asking an AI chatbot about a food today and receiving an answer based on information published several years ago. The answer might sound completely convincing. It might even have been correct once. But that doesn't necessarily mean it reflects the latest laboratory testing.
So, should you stop using ChatGPT for low FODMAP meals?
No! We actually think there's a much smarter way to use it. Think of AI as your creative kitchen assistant, rather than your FODMAP expert. Ask AI to:
✅ Come up with meal ideas
✅ Adapt recipes you've already checked
✅ Create shopping lists
✅ Help use up ingredients
✅ Suggest flavour combinations
✅ Make a recipe vegetarian or vegan
✅ Scale recipes up or down
✅ Help you batch cook
✅ Give you ideas when you're bored with your usual meals
But use a trusted source, such as the Monash App, to check:
🔎 Whether an ingredient is low FODMAP
🔎 The tested serving size
🔎 The relevant FODMAP subgroup so you can check for stacking
🔎 Whether the food has been retested recenly
🔎 Whether a packaged product has been independently tested
Here's a much better way to use AI
Rather than asking: "Create me a five-day low FODMAP meal plan." Try this:
"I have checked the Monash FODMAP App and confirmed that these foods are low FODMAP at these serving sizes: [insert your foods and quantities]. Please create five interesting meals using these ingredients. Do not determine FODMAP status yourself and don't add new ingredients without flagging them for me to check."
Now you're asking AI to do something it's very good at: being creative with information you've already verified.
You can even use it to adapt your favourite recipes. For example:
"Here's my favourite pasta recipe. I've identified the ingredients I need to replace because they aren't suitable for me. Can you suggest alternatives that will give me the same texture and flavour?"
That's a much safer and more useful role for AI for people following the low fodmap diet.
The smartest combination? AI + science
So, after almost two years of watching AI become dramatically better at answering questions, we remain convinced that you should use the Monash App for the science and use AI for the inspiration. That's probably the best way to get the benefits of both. And perhaps that's the biggest lesson from our original experiment. ChatGPT can give you a meal plan in seconds. But knowing whether that meal plan is genuinely low FODMAP? That's where you still need to do your homework.
A better way
If you want an easy way to try or re-try the low FODMAP diet, check out the 10 day meal plan in our Good Gut Challenge – we promise it includes reliably low FODMAP dishes that are way more exciting than the cheese and tomato on toast suggested by Chat GPT.
It includes our Monash certified low FODMAP shakes, which are an excellent source of low FODMAP protein, fibre, and probiotics.
A note about this blog
Our original blog on this subject was published in December 2024 and has been updated in August 2026 following a recently blog by Monash University. FODMAP information and recommendations can change as new foods are tested and existing foods are reviewed. Always check the current Monash FODMAP App for the most up-to-date food ratings and serving sizes and, for best results, consider working with a dietitian experienced in the low FODMAP diet.