According to Dongzha Beating Monitor, at the 2026 World Artificial Intelligence Conference, Xu Li, Chairman and CEO of SenseTime, delivered a keynote speech, elaborating on the dual proposition of inclusive and secure AI. Several AI applications that have already been market validated—such as software development, content generation, and embodied intelligence—are undergoing a common evolutionary path: transitioning from single-module tools for professionals to long-task closed loops for the general public. The most significant change in this process is the migration of users, where the original target customers are no longer crucial, and the real customers are emerging naturally from the demand side. Xu Li emphasized that true AI inclusivity is not about replacing humans but amplifying the inherent capabilities of each individual.

Xu Li also proposed that in terms of security governance, AI security should permeate product concepts, model pathways, and physical boundaries at three levels: products should not substitute but enhance, should not create dependencies but foster personal growth; models should ensure data security in both pre-training and post-training stages, and implement multi-model cross-validation during usage; embodied intelligence should gradually advance implementation through sandboxing and scenario constraints.

Xu Li further suggested that as the AI capability boundary advances to super-individual levels, the true future unit of account will no longer be a Token but the price of completing tasks—the Task Economy will drive AI from power consumption to value delivery, and continuous cost reduction will further lower the inclusivity threshold. Xu Li cited data from the World Economic Forum, stating that by 2030, AI might reduce 90 million jobs but create 170 million new professions, with super-individuals becoming the core of new occupational forms.

---------------------------------

Click the link below to access the original text and join the Dongzha Beating · Feishu AI News Channel, which continuously monitors global AI hotspots and news 24/7.