The system works like a search engine, but is designed to help chefs create brand new recipes based on ingredient and flavour pairings. Chefs have been serving up the system's suggestions in a touring IBM Food Truck to showcase the possibilities.
After the chef inputs a key ingredient into Watson, the program searches through its extensive database to find complementary flavours, before generating recipes based on compatible – but often unconventional – food pairings. This has led to unusual offerings like coconut-flavoured Caribbean Snapper Fish & Chips, Belgian Bacon Pudding, and the Austrian Chocolate Burrito, packed with ground beef and dark chocolate.
Mahmoud Naghshineh, vice-president of services research at IBM, said: "The idea of cognitive cooking is machines and humans working together – in this case, a very complex analytic system drawing from the vast collected knowledge of chemistry, food culture and taste preferences to help chefs break new ground."
The food truck was most recently seen at the South by South West film, music and tech festival in Austin, Texas this month. James Briscione, director of culinary development at ICE, said: "For a chef, creating flavour pairs is natural, but seeing ingredients matched in larger groups and understanding the science – the shared compounds that make these foods taste good together – is revolutionary."