At the 2026 Inclusion·Bund Conference, held from September 9 to 12 at the Shanghai Huangpu Expo Park under the theme "Co-creating the New AI Economy," renowned author Liu Zhenyun delivered a keynote speech. He argued that AI primarily grasps collective knowledge, which is essentially the aggregation of countless individual insights. However, he added a deeper layer to this idea, noting that AI's knowledge is limited to what currently exists or existed in the past.

Liu questioned whether AI will ever evolve to a stage of autonomous thinking. He suggested that to achieve autonomy, a system would first require its own soul, then develop emotions, and subsequently form a unique perspective on the world, an individualistic way of seeing things. Whether AI will ever reach this point remains uncertain, he said, and if that day were to arrive, the resulting relationship between AI and humanity would present a deeply compelling philosophical question.

"At this stage, when we talk about creativity, we mean producing something entirely new based on existing foundations—new angles, new thought processes, and new insights. From this perspective, current AI capabilities fall short," he stated. Liu cited related tests he has observed: instructing AI to produce a work modeled on A Feather in the Dust might yield titles like One Goose Feather or One Duck Feather; imitating I Am Not Pan Jinlian might generate I Am Not Ximen Qing; and One Sentence Is Worth Ten Thousand could be reversed into Ten Thousand Sentences Are Not Worth One. While AI can mimic these styles, the resulting text never reaches the literary heights of the originals.

He also emphasized that a work currently being conceived in a writer's mind but not yet written down cannot be imitated by AI at all. This proves, he concluded, that today's AI is merely an amalgamation of all historical and existing common knowledge, relying on computational power and big data rather than true originality.