News Releases

αU and Google's Gemini model to Provide Generative AI Services for Everyday Use

KDDI CORPORATION

Tokyo, February 21, 2024―From February 26, 2024, KDDI will launch an initiative toward providing services that enrich the lives of customers by enabling them to experience generative AI in their daily lives. To this aim, KDDI will combine its metaverse and Web3 service αU (Alpha-U) with generative AI, such as Google's high-performance generative AI model Gemini Pro model [Jump to the applicable section1].
As αU is boosting the creator economy, it is accelerating further support for creatives through generative AI.

The first services to be engaged will be shopping and entertainment. The former allows users to just take a picture of clothing and then check how clothes would appear without the users having to visit physical stores, by generating a try-on image. The latter allows users to generate live music scene in real-time. These services will be on show at MWC Barcelona 2024, with official launch planned for 2024 onward. Generative AI quickly creates wide-ranging creatives, which will allow customers to enjoy a personally optimized experience. To further broaden the use of AI, KDDI will open these AI functions to outside companies and provide development support.

In March 2023, KDDI and Google Cloud Japan partnered to co-create business by leveraging XR technology. Since then, Google Cloud's and KDDI's technical teams have been working together to create and disseminate new XR experiences. This time, the parties are extending the areas covered by this initiative to include the use of generative AI. As part of the development, KDDI will collaborate with the Google Partner Innovation Team, which conducts experiments using Google's state-of-the-art technologies to create innovative experiences based on generative AI.

If you have difficulty to view the movie, please click the link below.
Open link in a new windowαU(Alpha-U) and Google's Gemini model to Provide Generative AI Services

■Initiative detail

1. Enhancing αU with generative AI

  • KDDI operates the metaverse and Web3 service αU, including the metaverse, live streaming and virtual shopping, for new generations that sees no boundaries between reality and virtual space.
  • By linking αU, and generative AI such as Gemini Pro model, KDDI will support creatives and drive the creator economy.
  • Creative tasks will be made more easily, thereby allowing αU customers to experience more content and more enriching experiences.
  • Example use cases will be exhibited during MWC Barcelona 2024 at KDDI's booth. Official service launch is planned for 2024 onward.

<Examples of using generative AI>

Shopping:
Clothes fitting and customer service
  • Implemented in αU place, a shopping service that reproduces physical stores in virtual space.
  • From just one image of clothing, users can generate try-on images according to various body shapes. They can check how clothes would appear on themselves without visiting the store.
  • The AI chat service proposes items that are fit for the season and scene.
Live music streaming:
Venue in virtual space
  • Implemented in αU live, a 360-degree live music streaming service that can be viewed from any angle.
  • The live music venue switches to a forest, ocean, city, and other places in real-time according to the settings the customer enters by text.
  • Live music directors can quickly create a unique digital space.
  • Entertainment experiences can be customized according to each viewer.
Metaverse:
Creative Production
  • Implemented in a photographic Shibuya metaverse that is easily accessible with a Web browser through cloud rendering.
  • From just one creative graphic, countless digital townscape visions and posters in town can be automatically generated.
  • Creators can automatically generate consistent-looking creatives for any size and aspect ratio to effortlessly hijack the town, allowing customers to enjoy an immersive world-view experience in digital space.
Generative AI mascot:
Conversation engine
  • Implemented in generative AI mascot Open link in a new windowUbicot, which was announced as a prototype in October 2023.
  • Unlike smart speakers, it can talk emotively with generative AI.

2. Supporting companies by opening generative AI functions

  • KDDI will provide companies with functions that combine generative AI, such as Gemini Pro model with KDDI's technologies under the name of "αU on cloud".
  • Through Google Cloud Marketplace, companies can incorporate and enhance αU on cloud's generative AI functions in their products and services.
  • For startups, KDDI will offer additional support, such as matching them with major firms and offering development support, through KDDI's business co-creation platform KDDI ∞ Labo.
  • KDDI plans to start accepting applications from spring 2024.
  • [1]
    Gemini Pro model will be used at the MWC exhibit. Gemini Ultra model is to be supported when it becomes available in the future.
  • *The information contained in the articles is current at the time of publication.
    Products, service fees, service content and specifications, contact information, and other details are subject to change without notice.