*Subtitles: Choose “Subtitles/CC” in the “Settings” (gear icon) at the bottom right of the YouTube video.
*Subtitles may not show with multiple languages or overlapping audio.

In recent years, the emergence and rapid advancement of generative AI and large language models have taken the world by surprise. AI evolution is expected to significantly enhance productivity and drive scientific progress, allowing humanity to lead richer and more fulfilling lives. However, at the same time, concerns are growing over potential risks such as AI misuse leading to accidents, the militarization of AI, and even the possibility of AI running out of control.

This program aims to explore how humans and AI can coexist and move toward to "design future society for our lives." To delve into this critical question, we have invited three AI experts specializing in key areas such as autonomous AI agents, AI safety, and AI alignment to discuss the future of human-AI interaction.

【Reflection】
This session provided an opportunity to explore, from multiple perspectives, how humans can coexist with rapidly evolving AI technologies and shape the future of society and science. The moderator invited participants to consider how AI development may influence our ways of living and decision-making, prompting discussions across ethics, creativity, and democratic governance.

The session highlighted the possibility that, by 2050, a significant proportion of research may be conducted either autonomously by AI or collaboratively with AI systems. The concept of an “AI scientist” capable of independently generating hypotheses and drafting research papers was introduced, illustrating how AI could accelerate scientific exploration and creativity while humans assume roles in supervision and verification. It was also emphasised that broad trial and error, including failures, remains crucial to scientific advancement, and that AI has the potential to complement this process.

Attention was also given to the complexity of AI-related risks. These range from technical imperfections, privacy breaches, fraud, and child exploitation, to potential threats such as cyberattacks or military applications. The session highlighted that addressing these risks requires more than technological solutions, emphasising the importance of alignment with social systems and policy frameworks. At the same time, the need for flexible policy design and the careful use of societal simulations was underscored, in order to avoid opportunity losses due to excessive regulation.

The session further examined the gap between the rapid development of AI and society’s capacity for decision-making. Practical examples were presented, including citizen-participatory projects that demonstrate how AI and human collaboration can support democratic decision-making. The concept of “digital twins” was introduced to illustrate how individual values and preferences could be represented in AI systems. Such approaches offer the potential to enhance social inclusion and expand citizen participation, with AI serving as a supportive tool for democracy.

During the panel discussion, key themes included AI transparency and openness, balancing regulation with creativity, education for technical and policy professionals, and ensuring opportunities for societal participation. Overall, the session reinforced the idea that AI is not merely a technological challenge, but a co-creative partner capable of evolving with humans and contributing to society, encouraging participants to actively engage with the future.

【Post EXPO Initiatives】
Each speaker highlighted the following as key perspectives for their respective research and business development after the conference:
(1) Information Disclosure / Measuring Social Impact / Establishing Democratic Governance
To enhance transparency and accountability in AI development, promote the disclosure of model structures, training data, and evaluation criteria, and advance the development of frameworks for quantitatively measuring AI's social impact.
Furthermore, establish methods for democratically operating AI in practical settings such as policy formulation, citizen participation, and education, aiming for the “collective intelligence utilization” of AI.

(2) Human Resource Development and Building Cross-Sectoral Decision-Making Systems
Raise the baseline AI literacy of governments and companies worldwide. Promote human resource development and organizational building so that policymakers can understand AI and appropriately utilize and regulate it. Particular emphasis should be placed on designing a “hybrid decision-making system” where cross-disciplinary experts and democratic representatives collaborate.

(3) Building a Diverse AI Ecosystem / Expanding Experiential Opportunities
To address biases and risks in AI, promote the coexistence of diverse AI models independent of specific value systems. Reflect differing needs across technologies, cultures, and regions to build a truly diverse AI ecosystem. Furthermore, we will develop a cultural foundation for coexistence by expanding education, public dialogue, and experiential opportunities, enabling citizens to correctly understand both the benefits and risks of AI technology.

Through these efforts, we aim to establish AI as a “co-creative partner that extends human intelligence” within society, striving toward a sustainable future by 2050 where technology and humanity mutually enhance each other.

*Part or all of this report was generated by AI.