NABLAS Selected for GENIAC Phase 3: Advancing AI in Fact-Checking
NABLAS, an AI research institute based in Bunkyo-ku, Tokyo, has received exciting news regarding its participation in the third phase of the Generative AI Accelerator Challenge (GENIAC). This initiative, organized by Japan's Ministry of Economy, Trade and Industry and the New Energy and Industrial Technology Development Organization (NEDO), aims to enhance the country's capabilities in developing generative AI technologies.
This selection marks NABLAS's second consecutive acceptance into the GENIAC program, having previously been selected in the second phase last October. This latest project focuses on the research and development of a Japanese language-specific fact-checking large language model (LLM) and AI agents designed to combat the spread of misinformation within media and government institutions.
Objective of the Project
As misinformation continues to pose significant societal challenges, especially through the rapid dissemination of fake news, NABLAS aims to develop specialized LLMs that can effectively ascertain the veracity of information. The primary goals of this initiative include the creation of an autonomous fact-checking AI agent capable of both determining the truth of claims and generating reports based on those assessments.
To facilitate practical implementation, NABLAS has partnered with TV Asahi, a prominent Japanese broadcasting company, to launch real-world verifications and ensure the social application of the technology developed.
GENIAC: Empowering AI Development in Japan
The GENIAC program offers essential resources aimed at strengthening Japan's generative AI research capabilities. It provides computational power for building foundational models, supports data utilization, and assesses the effectiveness of AI applications. Through this initiative, NABLAS will work on the creation of its fact-checking LLM, leveraging reward modeling technology to build a robust and effective solution capable of addressing the challenges posed by misinformation in society.
Project Details
- - Project Name: Development Research of Factcheck RM and Agent
- - Implementation Period: August 2025 (date of approval) to February 2026
- - Focus Areas: Development of specialized models for fact-checking based on LLMs and creating an AI agent system for real-world application.
About NABLAS
Founded in March 2017, NABLAS is a venture spun out of the University of Tokyo, committed to fostering AI talent and providing groundbreaking AI solutions. The institute is dedicated to creating comprehensive educational programs geared towards training the next generation of AI professionals, utilizing updated content developed in collaboration with the university and presented through their iLect System learning environment.
In addition to its educational efforts, NABLAS offers consulting services for AI technology integration, research, and development tailored to meet the unique needs of its clients. With a mission to discover and innovate technologies that will lead to a better future, NABLAS is committed to the development of sustainable AI solutions that are beneficial for society.
For more information about NABLAS and its initiatives, visit
www.nablas.com.
Conclusion
NABLAS's participation in GENIAC Phase 3 emphasizes its commitment to utilizing AI technology for the betterment of society. By tackling the widespread issue of misinformation through innovative solutions, NABLAS is poised to contribute significantly to the enhancement of media reliability and public trust.