When you think of the classic burrito, you probably don’t think of Austrian cuisine. And you certainly don’t think of chocolate, apricots, oranges and vanilla beans. Which is why the Austrian chocolate burrito is so fascinating – it somehow manages to reinvent everything you thought you knew about which ingredients and flavors should be in a burrito.

The Austrian chocolate burrito is just one of the more than 65 recipes in the new Cognitive Computing with Chef Watson cookbook, a three-year collaboration project between IBM and the Institute of Culinary Education to showcase how supercomputers can be used in the kitchen to find new ingredient combinations and new ways of pairing foods.

Okay, the Austrian chocolate burrito may be creative and innovative, but is Chef Watson really capable of crafting a burrito that tastes better than what you’d find at the nearest Chipotle?

The answer, surprisingly, is yes.

The same Watson that beat two former grand champions on “Jeopardy!” appears also to be remarkably adept at discovering new recipes for chefs and home cooks. The recipe for the Austrian Chocolate Burrito was honed in the ICE kitchen and then field-tested using a food truck at Austin’s SXSW in 2014. The result is something special that is definitely a burrito, just not the kind of burrito you’d expect.

Let’s start with the ingredients. Instead of jalapenos, chili peppers or any other form of “heat,” there’s dark chocolate and ground cinnamon. Instead of red or black beans, there are green edamame beans. And instead of grated cheese, there are cheese curds (although using queso fresco or Cotija cheese is recommended in Chef Watson’s “Pro Tips”).

Then, there’s the matter of exactly how to combine the ingredients. If you’ve been to Chipotle, you know the drill: take a flour tortilla, pick your meat ingredient, pick your beans, and then start adding toppings such as cheese, onions, guacamole or sour cream. The result is then all wrapped up in the tortilla, and the ingredients blend beautifully together as you eat more and more of the burrito.

The assembly process was similar for Chef Watson, except that the meat gets mixed with chocolate, ground cinnamon and orange zest from the very beginning, giving it a browner, smokier taste. Then, the edamame beans are briefly boiled, washed and pulsed in a blender, giving them a slightly different texture. Finally, instead of adding sour cream or salsa at the end, there’s an apricot puree made with vanilla bean and dark chocolate.

#austrianchocolateburrito #burrito #chocolate #chefwatson #cognitivecooking #yum

A photo posted by @basulto on

The result was a dish that definitely boasted flavor complexity, even though it didn’t include any “hot” ingredients. And it was green, thanks to the edamame beans poking through, which made it look as though it had been overstuffed with guacamole. In an informal taste test of the burritos, the most common question asked was, “Is there really chocolate inside these?” (Which maybe explains why this recipe was off the charts on Chef Watson’s internal scale for “surprise.” Ha.)

So why was the Austrian Chocolate Burrito such a crowd-pleaser? The answer may have something to do with what the folks at IBM Watson refer to as “hedonic psychophysics,” or the study of what people find pleasant or unpleasant.

In layman’s terms, it means that your brain plays tricks on you when you taste something. So much of what we eat carries a lot of cultural and psychological baggage with it. When we think of Mexican cuisine, we naturally think of spicy flavors such as chipotle and we expect food to look and taste a certain way. Except that IBM Watson figured out that you could get similar pleasure sensations by combining flavor compounds found in Austrian cuisine and a creative use of chocolate.

In short, it means that you can have a Chipotle burrito without the chipotle.

And that was one of the primary goals of the Chef Watson project – to showcase the creativity, discovery and innovation that computers are capable of today. The key is that Chef Watson is not attempting to replace a chef – it is only acting as an aide and a consultant to make the cooking process more creative. For each recipe, Chef Watson provides a guide to three main factors – “surprise,” “pleasantness” and “synergy” – to help cooks choose the right offering.

What Chef Watson has done essentially is study the cookbooks from various cuisines, calculate which ingredients are typically paired with each other and which ingredients share underlying traits. It then cross-references this knowledge against a huge database of recipes to tease out unique suggestions of ingredients for a human chef to group together. You can see that in some of the other recipes suggested by Chef Watson. A classic French Canadian poutine, for example, becomes a Peruvian potato poutine with accents from the Andean highlands. Other offerings – such as the Thai-Jewish chicken and the Tanzanian-Lithuanian bagel – blend tastes from two completely distinct culinary traditions.

While many of the recipes are a bit too complex to recreate unless you have some serious cooking experience, Chapter Five of the cookbook is dedicated to “home cooking.” These recipes have been vetted by the folks at Bon Appétit magazine for people who enjoy challenging themselves at home, but who also want recipes that have been streamlined in terms of ingredients and time commitment – such as a spicy tomato gazpacho soup with ginger, designed to “surprise foodies from Seville to Bangalore.”

The big takeaway lesson is that, every now and then, it’s nice to spice up the classics with a creative, inventive twist. Hence, the phenomenal success of fusion cuisine and the relentless experimentation by restaurants (including fast food chains) to find new dishes and food combinations that will resonate with consumers. Eventually, super-smart computers will provide the creativity and innovation needed to create these dishes – the same creativity that’s historically been the exclusive preserve of top chefs and culinary experts.