Convenience store food in Japan inspired by AI

Fears that advanced artificial intelligence (AI) could wipe out humanity have dominated global headlines, but in Japan the technology is being used for a sweeter purpose — developing never-before-seen convenience store desserts.

For customers tired of the usual egg sandwiches and rice balls, major chains are turning to AI to develop products, including some unexpected flavors.

Lawson, one of the country’s biggest convenience store operators, yesterday said that it would roll out three new food items, including a pickle-flavored lemon tart and a bread roll filled with red bean paste, margarine and yoghurt cream.

People walk past the FamilyMart logo outside a store in Tokyo on Wednesday last week.

Photo: Reuters

The products, to be launched this month in Tokyo and neighboring regions, were created using AI around the theme: “What on earth is this? But it’s delicious.”

“To generate ideas unconstrained by existing ways of thinking or market trends, we used AI during brainstorming, without relying on past sales figures or market data,” the company said in a statement.

Meanwhile, rival FamilyMart said that it fed sales data into AI to help develop a new dessert — a French-style canele pastry made with dough containing sweet potato puree and topped with caramel sauce.

The dessert is to hit shelves nationwide next week, a spokeswoman told reporters, adding that the company had wanted to create something new, but likely to sell.

Like Lawson’s products, AI was used primarily for brainstorming, while human specialists were still involved in product development, the company said.

It is not FamilyMart’s first dessert product that was suggested by AI.

In July, the chain test-launched its first AI experiment — a biscuit sandwich filled with thick milk cream.

It proved a success, so the company decided to properly pursue AI in product development and plans to launch more of the items in the next few months, the spokeswoman said.

The rapidly advancing capabilities of AI have sparked concern in the past few days.

Researchers have warned of the risks of “superintelligent” computer systems that could evade human control.

Leading US AI labs have called for a coordinated slowdown of their cutting-edge work.