Chinese Q&A Platform Zhihu Releases AI Search Function

Pandaily03-22

Since the surge of large AI models over a year ago, the initial form of competition among domestic basic large models has emerged. The next step is to start competing at the application level.

In all large-scale consumer application scenarios, Internet content platforms are considered one of the most ideal directions for the landing of AI large models due to characteristics such as large user base, high usage frequency, and natural demand for generative content. Therefore, how content platforms utilize AI to transform themselves in order to better align with the content orientation of users in the AI era has become an important support for maintaining growth and innovation at this stage.

Established for thirteen years, Zhihu has become one of the indispensable original content platforms on the Chinese internet with a large group of professional creators and a high-quality Q&A community. Positioned in the most sensitive field to content innovation trends, Zhihu has undergone several revisions covering content formats, product features, etc., all to maintain its status as one of the most reliable communities in this era – a goal that remains unchanged even in the AI age.

On March 20th, Zhihu held the “2024 Zhihu Discovery Conference”, revealing the latest developments on the Zhihu large model and introducing a new AI feature called “Discovery·AI Search”, as well as many new initiatives in content mapping. This year, it hopes to become a more professional and trustworthy community, which will leave more room for imagination.

“When the community becomes a trusted home, there is an opportunity beyond just the community,” Zhihu founder, chairman and CEO Victor Zhou said.

More professional and reliable AI search

In the field of TOC products in AI, Zhihu may focus on AI search, as this highly aligns with the mindset and needs of its community users.

At the 2024 Zhihu Discovery Conference, Zhihu launched a new AI feature called “Discovery·AI Search.” According to Victor Zhou, this feature applies the “Zhihuitu AI” large model, using trusted content from the community as a source to provide users with a new experience including search, real-time Q&A, follow-up questions and other functions.

According to the introduction, Zhihu‘s ‘Discover·AI Search’ is a collaboration with Mianbi Intelligence’s large model, while the code work for search mainly adopts self-developed methods. Before the official release, Zhihu had tested it internally and on the platform for some time, and most of the feedback received by the team exceeded expectations.

In terms of manual annotation of the corpus, Zhihu has not yet made large-scale investments. The current effect is mainly based on the combination of wall-facing intelligent large model technology and Zhihu‘s accumulated high-quality content library. However, in the future, Zhihu stated that it will conduct specific annotations for fields such as science and finance to further optimize the presentation of search results.

In Zhihu‘s view, ‘Discover·AI Search’ has enhanced the convenience and usefulness of the community. At the same time, Zhihu is also continuously thinking and designing new feedback mechanisms for data presentation, creative distribution, and knowledge plan. Victor Zhou revealed that Zhihu will explore an open model with partners to allow creators to benefit while maintaining respect for and encouragement of original content.

“The truly valuable information is scarce.” Victor Zhou said that AI has magnified the value of the community, and the value of community interaction and discussion is unique to humans. Relying on a reliable community, the technological dividend brought by AI will empower Zhihu to continue evolving, achieving true “originating from the community, not limited to the community”.

SEE ALSO: Online Q&A Platform Zhihu Releases Large-scale Model

Looking at the community’s underlying color from the perspective of the new role ‘navigator’

At this Zhihu Discovery Conference, Zhihu announced the creative achievements of the first Lighthouse Program and launched ‘Haiyan Plan 6.0: Navigator’.

The ‘Haiyan Plan’ is a long-term support for creators constructed by Zhihu around dimensions such as creation system, traffic, and revenue. At this year’s Discovery Conference, Zhihu‘s COO, Senior Vice President, and Community Business Leader Zhang Ning announced the ‘Haiyan Plan 6.0’, introducing a new role – ‘Navigator’ to the community. Zhang Ning explained that while Haiyan Plan 2.0 introduced ‘Crowd Judges’ to help improve the discussion atmosphere in the community, Haiyan Plan 6.0 aims to introduce new roles through community interaction methods to present more good content.

According to the introduction, “Navigator” is a passionate user of the community, existing in various fields of the community. Their “approval” and “disapproval” will become high-value signals for Zhihu to distribute content and manage the community. The interaction of “Navigator” can help algorithms better understand opinion-based content, then expand to similar users of the same type, allowing “approval” and “disapproval” to generate greater value.

Zhihu‘s ‘Navigator’ status is implemented through an official invitation system, targeting two types of Zhihu users for invitations: one type is professional users who have completed the small blue badge identity verification in their respective fields; the other type is users with high activity in the community and representative consumption and interaction behaviors. In addition, Zhihu will also provide ‘Navigators’ with exclusive peripherals, priority product experiences, priority participation in offline activities, and other exclusive benefits.

Not just limited to the community

Obviously, for Zhihu, the content is not limited to just the community. ‘Zhihu started with a community. When the community becomes a trusted home for everyone, it has more opportunities than just being a community,’ Victor Zhou said.

Outside the community, Zhihu also has two major brands: Yan Story and Zhihu Zhixuetang. Yan Story is one of the achievements of Zhihu‘s ‘originating from the community, not limited to the community.’ Since 2019, Zhihu‘s paid reading business has gradually grown stronger and successfully pioneered a short story innovation model and track with Yan Story as its base.

The data disclosed at the conference shows that the number of Yan Story contributors has exceeded 600,000, with over 100,000 short stories published online and covering more than 180 subcategories. Zhihu also provides a channel for short story authors to monetize their creations, having distributed over 1 billion yuan in total earnings to creators and helping hundreds of selected authors achieve million-yuan incomes. In 2023, contracted authors had an average monthly income close to ten thousand yuan, ranking among the top in the industry.

In fact, Zhihu has already made a longer-term judgment on the short story track, believing that it will reach a scale of tens of billions in the next three to five years.

To this end, Zhihu has officially launched the “Short Story 3A Plan” driven by AI artificial intelligence, All Population and All Media. In terms of AI applications, Yan Story will explore the application innovation of large models in short stories in areas such as Text-to-Image, Text-to-Video, and creative assistance. It is introduced that large models have been widely used in Yan Story cover generation and will also be applied to Yan Story video generation scene.

Facing the AI era, Zhihu has already grasped the ticket of large models. In addition to the current clear search direction, there will be more opportunities in the future to try innovative applications of generative AI. When these revolutionary tools are combined with its community identity, perhaps a brand new chapter belonging to Zhihu will be opened.

“In the future, Zhihu will become an open loop and deepen its roots on a trustworthy basis while expanding more open new opportunities,” Victor Zhou said.

SEE ALSO: Zhihu Completes Homecoming IPO in Hong Kong

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

We need your insight to fill this gap
Leave a comment