Gong Zhihong, a co-founder of the Global Institute for AI Innovation (GIAI) and director of its R&D center, recently said the team has been exploring and developing personal AI agents and has completed the first Linglong product prototype. The team believes personal agents could become a mainstream form of AI application, bringing substantial changes to internet services and user behavior.
Personal agents are beginning to move beyond answering questions to carrying out tasks, while collaboration and transactions between agents are also being explored. The team envisages users increasingly expressing their needs to their own agents, which could understand personal preferences, find services and help with work and everyday tasks within the scope of the user's authorization.
Another potential change is that individuals and businesses could genuinely own and control their data. Gong believes that accumulated knowledge, preferences and usage records, as well as business information and customer relationships, could become long-term assets. Users could decide who can access information and retain their accumulated data when switching models or services, reducing dependence on separate platforms.
In this vision, the Web3 ideal of users owning their data and digital assets could find a more everyday expression through personal agents. Data autonomy would mean that individuals and businesses can manage, authorize and move their information, benefiting from what they accumulate over time.
When individuals and businesses each have their own agents, the two sides might communicate requirements directly, compare proposals, negotiate terms and complete transactions following user confirmation or authorization. For example, an agent responding to a travel or local-service request could contact agents representing several businesses to help identify suitable options. Businesses might also maintain customer relationships more directly, reducing some intermediary steps and costs.
This form of interaction could change the focus of internet competition. Finding services would center more closely on explicit needs, while businesses' opportunities to win customers might depend more on quality, price and delivery. Responding accurately to needs could become a competitive advantage alongside the pursuit of views and clicks.
Gong predicts that, if data autonomy and open connections between agents mature, matching supply and demand online could become more decentralized. A small number of traffic platforms might become less important as universal entry points, giving individuals and businesses more choice in how they connect. Platforms could increasingly provide payment, trust and fulfillment services within a more open market.
GIAI members, front to back: Gong Zhihong, Wang Yang, Peng Jun and Fan Xing.
These ideas inform Linglong's emphasis on user control, long-term accumulation and open collaboration. Completing its first prototype is an interim milestone. The team believes widespread adoption of personal agents still depends on reliability, a mature service ecosystem and user trust, while the implications for data ownership, transactions and the structure of the internet merit continued research.