Japan’s Rakuten plans to launch its own artificial intelligence language model within the next two months, its CEO told CNBC in an interview that aired Monday.
It comes as the fintech-to-e-commerce giant looks to join other technology firms developing the rapidly growing technology.
Hiroshi “Mickey” Mikitani said the company is working on its own large language model, or LLM. These are huge algorithms trained on massive data sets that underpin artificial intelligence applications, such as OpenAI’s ChatGPT.
Rakuten has a number of businesses from banking to e-commerce and telecommunications, therefore has a large amount of “very unique” data to train its LLM on, according to Mikitani.
“Nobody has a dataset like we do,” he added.
The company plans to use the AI model internally to improve operational efficiency and marketing by 20%, Mikitani said.
He also wants to offer the model to third-party businesses to build on, much like Amazon or Microsoft do.
“So we can easily teach them [businesses], package it and provide the platform for them to completely they can use it for their business,” Mikitani said.
The CEO added that Rakuten is going to “have something within a couple of months.”
To date, major U.S. and Chinese technology giants have been launching their own large language models.
OpenAI, Amazon and Google are among the most notable in the U.S. In China, Baidu, Alibaba and Tencent have launched their own models too.
Japanese firms have fallen somewhat behind their U.S. and Chinese counterparts. But they are trying to quickly catch up.
Telecommunications group NTT announced this month that its proprietary LLM will be available in March.
The telecommunications arm of SoftBank announced in November that its generative AI computing platform is operational.
Japanese firms have a chance to create LLMs specific to the Japanese language, potentially giving them an edge over their U.S. and Chinese rivals.
Mikitani said the push into AI is going to give Rakuten “huge profitable growth.”