People Shaping the Future of AI seriesHow Sovereignty Can Unlock AI Agent Collaboration

In the early 2000s, four or five years into his career at Fujitsu, Akira found himself working under intense pressure at the forefront of efforts to advance wireless communications technology. Mobile communications were at a turning point as the industry shifted from 2G to 3G. Driven by pride in Japan’s role as a global technology leader, Fujitsu and many other companies were working hard joined together to reach the world standard. Working alongside a leading telecommunications operator, the team was also striving to achieve world-first innovations. Expectations were high, schedules were relentlessly tight, and there was constant pressure to move faster. Spending the night at the office was not unusual. By today’s standards, it was an extraordinarily demanding working environment.

Even so, there was a genuine sense of excitement. Seeing the technology he had created become part of society's infrastructure and used in the real world left him with a lasting sense that technology could make a tangible contribution to people’s lives.

In 2012, Akira transferred to a research laboratory in San Jose, California. There he worked alongside researchers who had the freedom to pursue new ideas through their work, unconstrained by product requirements and supported by strong collaborations with universities. Adapting to a new language and culture was challenging, but the experience also introduced him to new ways of thinking. Through his work with multinational teams, he became deeply involved in information-centric networking, an emerging approach to network technology.

Unlike conventional networks, which communicate by specifying the location of a computer or server, information-centric networking is based on requesting and retrieving the data itself. It also includes mechanisms for verifying that the data has not been altered and can be trusted. This marked a fundamental shift from a location-centric to a data-centric approach to networking. The challenges Akira took on and the expertise he developed during his four years in the United States would eventually form the foundation of his work in AI Spaces*1, an advanced field of research.

*1 AI Spaces is a dynamic value-creation ecosystem where, upon a 'trusted digital foundation', multiple AIs autonomously learn and cooperate through linked data, transcending corporate and organizational boundaries.

Over the past decade, rapid advances in AI have transformed the value organizations can derive from their data. Yet the performance and accuracy of AI depend heavily on the quality and quantity of data available to it. When data remains confined within individual companies or organizations, its potential to generate innovation is less likely to occur. “I want to contribute to a society where collaboration among companies’ AIs continually generates new insights and allows people to live with greater convenience and peace of mind.” Driven by this ambition, Akira is exploring how technology can enable such collaboration while safeguarding sovereignty.

Turning that vision into reality requires technologies that allow organizations to interact with external AI without giving up control. But Akira believes technology alone is not enough. As with the early days of cloud computing, he sees one of the biggest challenges as helping organizations gain the confidence to take the first step. Demonstrating clear benefits, addressing concerns, and creating an environment where new approaches can be adopted responsibly are all part of the journey.
One such technology is the secure inter-agent gateway. Acting as a checkpoint between an organization and the outside world, it monitors and controls real-time information and actions its AI agents attempt to send externally in real time. Policy control technology mathematically models a company’s transaction policies and governance rules, teaching AI agents what they may and may not do. Guardrail technology helps secure interactions between AI agents by preventing confidential information from being inferred or exposed.
Together, these technologies form part of a framework designed to give organizations the confidence to entrust AI agents with greater responsibility. Akira often describes this challenge as requiring “first penguins” – those willing to take the initial risk and explore new possibilities.