AI图景下大数据挖掘风险评估与应对策略.doc

AI图景下大数据挖掘风险评估与应对策略 　　〔摘要〕[目的]揭示人工智能（AI）迭代发展中大数据挖掘的风险、成因与对策，为打造智能化协同创新生态链提供重要支撑。[方法]分类探讨AI图景下大数据挖掘中智能采集侵害数据产权、关联勾勒突破数据匿名、黑盒疏漏打破动态平等，以及算法决策冲击自由抉择等风险及其成因，提出数据挖掘的原则与策略。[结果/结论]参与者之间参差不齐的管控意识与处理能力、复杂多元的利益诉求以及缺少政策法规集中调控等导致上述风险，亟待严格遵循平等参与、优质输入、彻底拭除以及公允运转等挖掘原则，主动采取充实数据产权规定、丰富数据隐私条款、明确挖掘主体资质与公权参与标准、推动签署数据跨境统一处置方案以及健全相关救济措施等具体对策。 　　〔关键词〕AI图景；大数据挖掘；算法决策；风险评估 　　DOI：10.3969/j.issn.1008-0821.2018.05.023 　　〔中图分类号〕D92217〔文献标识码〕A〔文章编号〕1008-0821（2018）05-0147-05 　　〔Abstract〕[Purpose]The paper exposed risks，causes，and countermeasures of big data mining in the iterative development of artificial intelligence.It would help to build an ecology chain of intellectualized synergistic innovation.[Methods]The article classified many risks in big data mining，which included infringing data property in the course of intellectualized data collection，changing data anonymization in drawing the outline of relevance，disrupting dynamic equality by the shortage and loss of the black box，striking the free will by algorithm decision-making mechanism，and suggested data mining principles and countermeasures.[Results]Different controlling ideas and varied management abilities of participants，complicated and diversified interest requirements，and the correspondent lagging policies and laws，brought above risks.It was necessary to abide by big data principles，such as equal participation，high quality inputs，complete deletion，and fair running；enrich provisions for data property right，improve rules of data privacy，define the mining qualification and the entrance criteria for public powers，promote the integrative international scheme about cross-border data mining，and perfect relative remedies，etc. 　　〔Key words〕AI picture；big data mining；algorithm decision-making；risk assessment 　　基于全球新一?科技革命和产业变革中最先大规模应用人工智能的国家将掌握竞争主导权和事务话语权的合理预期，我国从“互联网+人工智能”、“新一代人工智能发展规划”到十九大报告中强调“推动人工智能和实体经济深度融合”[1]以及十三届全国人大一次会议政府工作报告提出“做大做强新兴产业集群，实施大数据发展行动，加强新一代人工智能研发应用”[2]，愈加重视人工智能研发与应用的战略部署。庞大的网民基数、高速发展的基础设施和万物互联的节点布局等为我国带来得天独厚的数据优势。执