AI’s Potential Is Being Curbed by Lack of Trust and Talent, Experts Say

The artificial intelligence (AI) industry is in its nascent stages but promises large investment opportunities, particularly in areas like humanoid robots and AI agents. Despite robust interest, the sector faces obstacles such as integrating data, building trust, and overcoming a shortage of skilled professionals, as discussed at the Asia New Vision Forum 2025. Experts note that these factors currently hamper widespread adoption, suggesting that the industry remains far from reaching its full potential [para. 1].

Xiong Youjun, general manager of the Beijing Innovation Center of Humanoid Robotics, pointed out that although there has been a surge in interest, robot investments in China remain modest compared to North American counterparts: only tens of billions of yuan have been invested so far. He emphasized that humanoid robotics is still in an early stage, with current investments seen as only a precursor to future developments. As several companies begin small-scale production, mass-market entry and the corresponding need for larger investments in critical areas are expected next year. The long-term potential is immense, with Xiong predicting that once humanoid robots become household items, the market could be valued in the trillions of US dollars [para. 2][para. 3][para. 4][para. 5][para. 6].

Supporting this optimism, Zhao Yuli, chief strategy officer at Galbot, referenced a Morgan Stanley report forecasting that China's humanoid robot market will grow to $40 billion by 2035, a dramatic increase from its current scale. She emphasized that strong growth momentum exists, and success will come to those firms that best connect supply with genuine demand by transforming technology into scalable, industrial products [para. 7][para. 8].

On the topic of AI agents, Cynthia Zhang, founder of FutureX Capital, noted that while AI capabilities are ready, there remains a lack of compelling products, especially in the AI agent space. Major tech companies worldwide, including Google, Microsoft, Anthropic, Baidu, Lenovo, and Zhipu AI, have entered this arena, targeting both enterprises and consumers. A key milestone was the launch of Manus, a general-purpose AI agent developed by Beijing Butterfly Effect Technology, which can autonomously perform intricate tasks like resume filtering and stock analysis [para. 10][para. 11][para. 12][para. 13][para. 14].

Zhang highlighted industry challenges, such as the lack of experienced engineering teams and the difficulty startups face in securing high-quality, domain-specific data and enterprise trust for collaborative development. Despite these barriers, she sees abundant investment possibilities in the next 5 to 10 years. Her fund has invested in 60 AI companies across China and the US, with 70% already profitable, stressing the importance of betting on capable founders rather than specific market segments [para. 15][para. 16][para. 17].

Hans Dekkers of IBM Asia Pacific added that there is space for multiple large language models (LLMs) and a diverse range of domain-specific models, especially those based on proprietary enterprise data [para. 18].

Despite overall optimism, the road to AI's broad adoption is hindered by issues like trust, safety, and data integration challenges. Panelists highlighted the need to build public confidence in AI, manage fears of job displacement, and demonstrate superior safety, particularly in autonomous driving. Dekkers identified significant potential in integrating disparate AI agents across organizations into unified platforms. In sum, speakers agreed that while the AI industry is just beginning to “explode,” its transformative impact will depend on which companies successfully lead that change [para. 21][para. 22][para. 23][para. 24][para. 25][para. 26].