HowNet "accused" the AI ​​search company Secret Tower of infringement, demanding that its website link be disconnected, Secret...

The core of the dispute between HowNet and the AI search company Secret Tower centers on the legal and commercial implications of large-scale data scraping for AI training and service provision. HowNet, as a major aggregator of Chinese academic resources, holds proprietary rights over a vast database of journal articles, theses, and conference papers. Its accusation against Secret Tower likely alleges that the latter's AI search engine, which presumably offers academic content summarization or retrieval, was built or trained using HowNet's copyrighted content without proper licensing or authorization. The specific demand to disconnect the website link is a direct and severe legal tactic, often a precursor to or component of a copyright infringement lawsuit, intended to immediately cease the alleged unauthorized dissemination or use of the proprietary data.

From a technical and operational perspective, Secret Tower's business model as an AI search company inherently relies on ingesting massive corpora of text to train its models and populate its search indices. The high-quality, structured, and domain-specific nature of academic literature makes it exceptionally valuable for training specialized AI. The alleged infringement, therefore, touches upon the unresolved legal gray area of whether using copyrighted materials for AI training constitutes fair use or requires explicit permission. HowNet's action suggests it views such use as a direct commercial threat, potentially undermining its subscription-based revenue model if AI agents can effectively parse and deliver its core content without users needing direct access to the HowNet platform itself.

The broader implications of this case are significant for China's burgeoning AI industry and its intellectual property landscape. It represents a high-profile clash between established data gatekeepers and new AI-driven intermediaries. A ruling or settlement favoring HowNet could force AI companies to undertake costly and complex licensing negotiations for all training data, potentially slowing innovation and entrenching the market power of existing database holders. Conversely, a precedent that broadly permits such data use for AI could destabilize the business models of content aggregators. The demand for a link disconnection also highlights the powerful role of network infrastructure and internet service providers in enforcing intellectual property claims, as compliance would effectively erase Secret Tower's public access to the contested content.

Ultimately, this accusation is less about a single link and more about defining the boundaries of data ownership in the AI era. The outcome will hinge on interpretations of copyright law as applied to machine learning processes and the extent to which the output of Secret Tower's service is seen as a derivative work or a transformative fair use. It serves as a critical test case, with the potential to establish norms that will either facilitate AI development by granting broader data access or protect data compilers by requiring formal consent for commercial AI utilization, thereby shaping the competitive dynamics of both the AI and academic information services sectors in China.

References

Related Questions