AI creates cookbook with horrifying results
AI creates cookbook, AI makes a lot of really horrible sounding recipes that confirm that AI don’t really know how food work.
There’s an endless debate raging involving the future of artificial intelligence, and how it will change the world. At this point, it’s pretty much inevitable that we will have fully functional AI, probably sooner rather than later, it’s more-or-less a given. And even if it is completely dedicated to helping humanity (which is a big “if”), there will still be a negative impact on society, even if the pros outweigh the cons.
Basically, it’s a question of whether AI will destroy us or just take all our jobs. Think about that the next time you ask Siri to schedule your meetings.
Regardless of how deeply the introduction of AI screws us, there are some fields that, logically, you’d think would remain the province of humanity. Take, for example, cooking. An AI can mimic a recipe, but without the necessary equipment – namely taste buds – it is still basically just guessing as to how it will all come out. But that hasn’t stopped people from trying.
Research scientist Janelle Shane decided to push the borders of sanity and teach a neural network to create recipes. Shane has earned a reputation for designing some bizarre and often hilarious scientific experiments on her Tumblr page, and this is just the most recent in a long line.
The experiment began by showing the neural network several recipes to give it an idea of where to begin. Shane also controlled the creativity quotient, allowing her to adjust and determine how imaginative the recipes should get. When set to low, the AI makes some simple guesses as to recipe titles it thinks people want. Here are a few at the low setting.
Cream Cheese Soup
Cream Of Sour Cream Cheese Soup
Chocolate Cake (Chocolate Cake)
Chocolate Chocolate Chocolate Cake
Chocolate Chicken Chicken Cake
Chocolate Chocolate Chocolate Chocolate Cake
Chocolate Chips With Chocolate Chips
You really have to wonder about the “Chocolate Chicken Chicken Cake.” Now here are a few at the higher setting.
Beef Soup With Swamp Peef And Cheese
Chocolate Chops & Chocolate Chips
Crimm Grunk Garlic Cleas
Export Bean Spoons In Pie-Shell, Top If Spoon and Whip The Mustard
Chocolate Pickle Sauce
Whole Chicken Cookies
Salmon Beef Style Chicken Bottom
Out Of Meat
Completely Meat Circle
Completely Meat Chocolate Pie
Cabbage Pot Cookies
Artichoke Gelatin Dogs
Crockpot Cold Water
And who wouldn’t want a hot, steaming cup of “Crockpot Cold Water,” or a nourishing plate of “Whole Chicken Cookies?”
When it broke down the recipes, things got even more horrifying. This dish is known as “Immediately Cares, Heavy Mim.” Delicious
Immediately Cares, Heavy Mim
3 dill loasted substetcant
1 cubed chopped whipped cream
3 unpreased, stock; prepared; in season
3 cup milk
1 ½ cup mOyzanel chopped
½ teaspoon lemon juice
1 ¼ teaspoon chili powder
2 tablespoon dijon stem – minced
30 dates afrester beater remaining
Bake until juice. Brush from the potato sauce: Lightly butter into the viscin. Cook combine water. Source: 0 25 seconds; transfer a madium in orenge cinnamon with electres if the based, make drained off tala whili; or chicken to well. Sprinkle over skin greased with a boiling bowl. Toast the bread spritkries.
Yield: 6 servings
To be fair, this was an early recipe from the almost-AI. As it learned, the recipes got slightly better – or at least they made a little more sense. The first recipes were little more than random letters and gibberish. It’s actually somewhat impressive that the neural net can piece together even this much, given that it can only process 65 characters at a time. That means that by the time it gets to creating the recipe, it has already forgotten what the name of the dish is.
Shane’s experiments were inspired by another culinary arsonist named Tom Brewe, who ran a similar experiment with equally tragic results. The recipes he produced included such appetizing sounding dishes as “Green Baked Stuff,” “Hot Garlic Casserole,” and “Basil Balls of Bean Bread.”
Below is the recipe for “Caramel Corn Garlic Beef.”
Categories: Soups, Desserts
Yield: 10 Servings
2 tb Parmesan cheese, ground
1/4 ts Ground cloves
1 ts Cayenne pepper
Cook it with the batter. Set aside to cool. Remove the peanut oil in a small saucepan and pour into the margarine until they are soft. Stir in a a mixer (dough). Add the chestnuts, beaten egg whites, oil, and salt and brown sugar and sugar; stir onto the boqtly brown it.
The recipe from an oiled by fried and can. Beans, by Judil Cookbook, Source: Pintore, October, by Chocolates, Breammons of Jozen, Empt.com
So get ahold of a neural network and feed it recipes! Have a dinner party and serve your guests the results! Make sure to invite people you really don’t like!
There are already plenty of companies looking to introduce robots that can cook, but those are essentially tools for chefs. People still need to buy the ingredients, tell the robot exactly what to do and how to do it, and hope that it follows the instructions precisely. It isn’t adding its own, unique flair to the meal and it can’t check to ensure that the meal tastes good.
But sooner or later neural networks will probably find an acceptable level of culinary acumen. They probably won’t be able to surpass any skilled chefs in creating a new dish or bringing the flavors out of ingredients, but they could be good enough for feeding huge numbers of people where taste isn’t the primary goal, like prison cafeterias or Chilis.
So think on that the next time you hit the McDonald’s drive-thru and order your nuggets with Szechuan sauce. One day soon, you may be ordering a new meal dreamed up by an AI.