人工智慧、個資及資料之可能法律議題：以擷取運動影像資料為例

Potential legal issues concerning artificial intelligence, personal data, and data collection: A case study on extracting sports image and video data

緒論：透過資料分析預測運動表現，或結合人工智慧（AI）進行影像辨識，在運動領域十分常見。我國雖有個人資料保護法或相關法令，卻無AI應用之專法，相關應用之合規性值得探討，且國際奧會與各運動管理組織亦未訂定自律要求。方法：有鑑於先前已有許多研究，討論大數據或其他在運動產業之資料運用，但仍欠缺對AI應用之研析，特別是運動與法律方面。故本文以運動影像資料為應用場景，並以經當事人同意之訓練資料蒐集為主、排除公開情蒐，先透過2018年英國法院對人臉辨識之重要案例研析與回顧，輔以歐盟人工智慧法與國內相關法令之研析，研提法令遵循與管理之應注意事項。結果：初步建議依我國現行AI之應用，係以訂定參考指引先行，建議運動產業之主管機關，可針對運動產業需求，訂定生成式AI或其他AI應用之參考指引。結論：除先於契約內明訂特定目的、利用期限等約定，以符相關應用之法令遵循外，未來可參考憲法法庭於相關判決之意旨，建議主管機關可進一步訂定個資保護或資料管理規範，以降低AI應用之可能風險。

Introduction: The use of data analysis to predict athletic performance and the integration of artificial intelligence (AI) for tasks such as image recognition has become widespread in the sports industry. Although Taiwan have enacted the Personal Data Protection Act and associated regulations, there is no specific legal framework governing the application of AI technologies in this context. This raises important questions regarding the compliance of AI applications with existing laws. Furthermore, the International Olympic Committee (IOC) and various sports governing bodies have yet to establish specific self-regulatory standards for AI use in sports. Methods: While previous studies have extensively addressed the use of big data and other forms of data processing in the sports industry, there is a notable lack of research examining the legal implications of AI applications, particularly in the domain of sports law. This article focuses on the application of AI to sports-related image data, and emphasizes that training data should only be collected with the consent of the individuals involved, and that data obtained through public surveillance should not be used. The analysis begins with a review of a landmark 2018 UK court case concerning facial recognition technology, and is further supported by an examination of the EU Artificial Intelligence Act and relevant domestic legislation. Key considerations for legal compliance and risk management in AI deployment within the sports industry are outlined. Results: The primary recommendation is the development of provisional guidelines for AI applications that reflect current industry practices and legal standards. It is proposed that regulatory authorities overseeing the sports industry introduce guidelines specifically tailored to address the use of generative AI and other emerging AI technologies. Conclusion: To mitigate the risks associated with AI applications in sports and ensure compliance with both domestic and international legal standards, contracts should clearly stipulate the specific purposes for which AI will be used, the usage limitations, and the data retention periods. Based on the relevant Constitutional Court rulings, it is also advised that regulatory bodies enhance the legal framework surrounding personal data protection and data governance in the context of AI.