Securities Times front-page commentary: AI is an assistant, not a stand-in; analysts remain the responsible parties for research reports.

Title context: Text: Generative AI has rapidly gained traction in the securities research field over the past two years. From draft research reports and meeting minutes to the construction of valuation models, AI has become deeply embedded. Along with efficiency gains, compliance risks have followed: some analysts have introduced AI-generated data and opinions directly into research reports without verification; worse still, AI-fabricated policy information and invented industry data have been signed off by analysts and sent to the market, causing problems such as policy timeline mismatches and untraceable data sources in research reports. The Securities Association of China recently issued the "2025 Statistical Report on the Business Operations of Securities Companies Publishing Securities Research Reports," which for the first time writes "establishing an internal control mechanism for reviewing AI-generated research content" into industry standards, explicitly prohibiting analysts from directly adopting unverified data, opinions, deductions, and research conclusions generated by AI. Although brief in wording, this provision carries substantial weight. It redefines the position of AI and analysts in the research chain: AI can be an auxiliary tool, but it cannot replace professional judgment; analysts can leverage technology, but they must be responsible for the final published reports. Technology itself is neutral. It can amplify research efficiency, but it can also amplify research errors. Only by putting AI's auxiliary role in its proper place, firmly establishing analysts' gatekeeping responsibility, and building a complete institutional internal control loop can the technology dividend truly research quality.