Perplexity AI Co-hosts Japan-U.S. AI Governance Symposium with Stanford University, the U.S.-Japan Foundation, and Kyoto University

Perplexity AI, Inc., the company behind the AI answer engine Perplexity, co-hosted the Japan-U.S. AI Governance Symposium on May 11, 2026, at SHIBUYA QWS together with Stanford University, the U.S.-Japan Foundation, and Kyoto University. The symposium brought together policymakers, legal scholars, technology leaders, and practitioners for in-depth, in-person discussions on the future of AI governance. The debate has clearly shifted from abstract principles to practical questions about AI adoption in business and government. Across the sessions, trust emerged as the central theme. This refers not only to trust in system performance, but also to trust in the broader process by which AI is governed and integrated into society. Effective governance requires coordinated action across government, industry, and the wider ecosystem. In his keynote address, Japanese House of Representatives member Taira Masaaki described AI as a national priority directly tied to economic competitiveness and national security. He said Japan will continue pursuing agile soft-law governance rather than EU-style comprehensive regulation, with the goal of becoming “the world’s most AI-friendly country.” He also emphasized the need to transform government systems themselves into AI-driven systems in anticipation of agentic AI. Japan’s concept of “AI sovereignty,” he said, is evolving beyond domestic model development toward strategic control over infrastructure, data, and cultural adaptation, combining international partnerships with domestic capabilities. Japan is also considering targeted enforcement measures, including penalties, to address threats such as cognitive warfare and risks to critical infrastructure. Panel discussions highlighted that while regulatory approaches differ by cultural context, including Japan’s flexible agile model, Europe’s legal-regulatory model, and the U.S. market-driven model, global approaches are increasingly converging toward Japan’s flexible, trust-based model. Japan is establishing an important position as a bridge among global regulatory approaches. On intellectual property, participants discussed the need for both litigation to clarify legal boundaries and practical cooperation through contracts, such as direct licensing and data-sharing infrastructure. RAG systems that clearly cite sources were identified as a realistic solution for enabling proper attribution and compensation. On social implementation, Japan’s AI adoption is at a turning point from proof-of-concept projects to large-scale deployment. In organizations that embed AI across workflows, the human role will shift toward oversight, judgment, and strategic decision-making. Hybrid architectures combining local and cloud systems will be crucial for ensuring trust and privacy. The symposium concluded that AI governance will be shaped by the interaction of policy, technology, business practices, and social expectations. Rather than relying on a single actor or regulatory approach, coordinated action among governments, AI providers, and the wider ecosystem is the path toward a trusted AI society. The symposium was held under the Chatham House Rule and featured candid discussion among leaders from Japan and the United States. Speakers included keynote speaker Taira Masaaki, as well as representatives from Japan’s Cabinet Office, GPAI, Tohoku University, Kyoto University, Waseda University, Mori Hamada & Matsumoto, Rakuten Group, Helios, Sakana AI, Microsoft Japan, and Perplexity. The event took place on May 11, 2026, at SHIBUYA QWS Scramble Hall. It was hosted by Stanford University and the U.S.-Japan Foundation, with support and cooperation from Kyoto University and Perplexity AI. Perplexity is an AI company building products and services based on accurate AI. Founded in 2022, its mission is to “power the world’s curiosity.” The company develops Perplexity, an answer engine that responds to questions with inline citations based on trusted sources and deep research. It also offers Comet Browser, an AI-native web browser equipped with the AI agent Comet Assistant, and is developing Perplexity Compute, a large-scale multi-model orchestration system spanning tools, files, code generation, persistent memory, and the open web. Perplexity currently answers more than 1.5 billion questions per month worldwide.