Japan – NEC, Kimuraya and ABEMA use AI to develop bread with a “taste of love”

NEC Corporation (NEC; TSE: 6701), together with Kimuraya Sohonten Co., a long-established bread maker, and Abema TV, a video streaming service, have teamed up to produce bread with flavors inspired by the romantic Japanese television program, “I Fell in Love with you Today.” NEC’s artificial intelligence (AI) has analyzed conversations from the cast of the show and the lyrics of songs that feature fruits and sweets (*1) to produce five varieties of bread, “Ren AI Pan,” that aim to link flavors with romantic feelings. Kimuraya plans to begin providing Ren AI Pan at supermarkets and online stores in areas surrounding Tokyo from February 1, 2024.

Kimuraya, was established more than 150 years ago and has traditionally marketed its products through sales channels such as department stores and supermarkets. As a result, acquiring new customers and recognition among younger consumers has been a challenge. Therefore, the company decided to team up with NEC, which has a track record of developing specialized food products using AI, including pudding and craft beer.

A recent market survey of young people in Japan, the target audience for this project, indicated that they are becoming further removed from romantic relationships. With this in mind, the partners developed a product that utilizes the content of a much-talked-about TV program, “I Fell in Love with you Today,” which features high school students and the dating issues they face.

1. “NEC Enhanced Speech Analysis,” which utilizes speech recognition technology and belongs to NEC’s AI technology group “NEC the WISE” (*2), was used to extract and analyze portions of conversation data from three seasons of “I Fell in Love with you Today,” including the Cebu, Phuket, and Okinawa editions of the program.
2. “NEC Data Enrichment,” which utilizes technology for understanding the meaning of data and belongs to “NEC the WISE,” was used to assign emotional scores to conversations from romantic scenes, such first dates, to visualize the emotion trends.
3. Approximately 35,000 songs that include 183 types of fruits and sweets and other foods were extracted from a database of approximately 1 million Japanese songs. “NEC Data Enrichment” was then used to assign emotional scores to lyrics containing food and to visualize the emotional trends associated with food images.
4. The romantic scenes and foods from (2) and (3) that have similar emotional tendencies were linked, and a list of the top 50 foods that express the emotion of love was created.
5. Kimuraya reproduced “love as a taste” by selecting a combination of foods that go well together.
6. Utilizing an NEC-developed Large Language Model (LLM), product names and descriptions for use in packaging and special websites were created, such as “Fate,” “First Date,” “Unforgettable,” “Tears of Heartbreak” and “Mutual Love.”