#Japan #Tokyo #NEDO #ABEJA #CODAS

Collaboration Between ABEJA and CODAS

ABEJA Inc., a Tokyo-based company aiming to implement a "rich world" through the collaboration of humans and AI, has officially joined forces with the Co-creation Organization for Data Sovereignty & AI (CODAS) under the National Research and Development Agency NEDO's initiative. This groundbreaking project, titled "Development of Efficient AI Learning Methods with Consideration for Data Confidentiality," focuses on establishing a robust framework that allows enterprises to utilize industrial data securely without worrying about security risks.

The Objective of the Collaboration

The primary goal of this initiative is to develop a "Privacy-First AI" that guarantees secure usage of industrial data, hence alleviating fears related to data breaches or privacy violations, which are concerns that approximately 70% of companies have in current scenarios. This innovative endeavor addresses security risks while creating high-performance AI solutions tailored to various industries. ABEJA has been chosen as a key partner in this project, particularly responsible for establishing a "Sector-Specific Large Language Model (LLM)" within the effort led by CODAS.

ABEJA's Business Vision

ABEJA operates with a mission to turn this vision into reality through its core platform, the ABEJA Platform. This enterprise platform successfully transitions AI from mere proof of concept into practical operations, addressing various aspects such as decision-making and operational execution by integrating Agentic AI and Physical AI. The platform fundamentally transforms data handling, decision-making processes, and operational frameworks to elevate business practices in real-life environments. ABEJA's architecture is based on the “Human in the Loop” principle, which ensures that AI is operational from the initial stages, eliminating the traditional Proof of Concept (PoC) phase, commonly known as “Zero PoC.” As a result, critical mission-oriented tasks can increasingly incorporate AI into their operations over time, enhancing efficiency and security.

CODAS and Its Objectives

The Co-creation Organization for Data Sovereignty & AI (CODAS) is a nonprofit entity aiming to establish a safe AI society by focusing on research and development of data privacy technologies and their subsequent open-source implementation. The urgency in their work stems from the fast adoption of LLM usage across various industries, yet a considerable amount of companies remain hesitant to adopt these solutions fully due to prevalent security concerns surrounding confidential information and the risks of losing data sovereignty. Therefore, this project aims to lay the foundation for "Privacy-First AI," providing a secure foundation for enterprises to manage and operate their data safely domestically.

ABEJA's Role in Ensuring Data Privacy

ABEJA's participation in this pivotal project is based on its proven expertise in LLM and associated technologies. The initiative is fundamentally structured around two pillars: the research and development of data anonymization technologies, coupled with the establishment of sector-specific LLMs. The technologies and models developed during this endeavor will be progressively made available as open source, facilitating the creation of an ecosystem where domestic companies can comfortably utilize these privacy-first AI solutions.

Initially focusing on the financial and insurance sectors, ABEJA plans to gradually expand its efforts to other foundational industries, such as manufacturing. They intend to rely on state-of-the-art security technologies, including Confidential Computing, which allows for cryptographic protection of data processing within secure environments. Furthermore, the integration of NVIDIA’s latest model, the "NVIDIA Nemotron 3", is anticipated to enhance the capabilities of the ABEJA Platform further, bringing high processing efficiency while meeting demanding security requirements.

Future Prospects

Looking forward, ABEJA is committed to continuing its role in assisting enterprises in employing AI effectively within mission-critical sectors, contributing to the broader transformation of industrial structures. This collaboration is expected to shape the future of data utilization securely and efficiently, thereby establishing a new standard for AI integration in diverse industries.

Conclusion

With organizations like CODE and ABEJA leading the charge, the landscape of AI application on a commercial level is about to shift drastically, providing a safer environment for data handling and fostering innovation across various sectors. The journey toward achieving a robust framework for Privacy-First AI is underway, and its success will undoubtedly pave the way for more secure technological advancements in the future.