生成式人工智能数据获取风险与规制路径
Generative AI Data Collection Risks and Regulatory Paths

生成式人工智能技术的应用对于社会发展和技术进步具有重大作用，但与此同时，由于技术限制、缺少具体应用规范等产生诸多风险。在生成式人工智能技术的运行中，数据获取是其基础也是风险源头，因数据来源多样，收集数据缺少明确的授权容易产生违规风险；不实信息、价值观偏差、文化单一等产生收集数据质量风险；海量数据传输过程中产生数据泄漏风险。为了防范、应对生成式人工智能数据收集引发的数据风险，应当确立从细化人工智能立法、到明确数据收集阶段各参与主体的职责、再到强化行政监管的风险规制路径。
The application of Generative Artificial Intelligence (GAI) technology plays a significant role in the development of society and technological progress, but at the same time, it generates many risks due to technological limitations and lack of specific application specifications. In the operation of generative artificial intelligence technology, data collection is the basis and source of risk, due to the variety of data sources, the lack of clear authorization to collect data is prone to violation of the risk; inaccurate information, value bias, cultural singularity and so on, resulting in the quality of the collected data risk; data leakage risk in the process of massive data transmission. In order to prevent and cope with the data risks caused by generative AI data collection, a risk regulation path should be established from refining AI legislation, to clarifying the responsibilities of each participant in the data collection stage, to strengthening administrative supervision.