AsianFin -- Jiang Daxin, founder and CEO of StepAI, outlined his vision for the evolution of AI agents at the 2025 Inclusion Bund Conference. He pointed out three major trends he believes will shape the future of intelligent agents.

First, Jiang predicts that agents’ effective working hours will continue to increase. As AI systems become more capable, they will be able to handle longer and more complex tasks without human intervention, effectively extending the productive time of enterprises and individuals.

Second, he foresees a transition from the digital world to the physical world. While current AI agents primarily operate in digital spaces—processing data, generating content, or assisting with virtual tasks—future agents will increasingly interact with the physical environment. This evolution will enable AI systems to bridge the gap between online intelligence and real-world action, integrating robotics, IoT devices, and other hardware platforms.

Third, Jiang anticipates a shift from experience-based learning to self-evolution. Traditional AI models largely rely on historical human-generated data, which is both limited and prone to bias. Jiang argues that to surpass human intelligence, AI agents must learn autonomously through interactions with themselves and their environments. This approach allows continuous adaptation, improvement, and the emergence of capabilities beyond what is encoded in the training data.