人工智慧發展下資料運用與個人資料保護之研究-以自駕車為例

A Study on Data Application and Personal Data Protection in Artificial Intelligence Application -Take the Automated Vehicles as the example

黃國忠(Kuo-Chung Huang)

指導教授 ： 黃銘傑

國立臺灣大學/進修推廣學院/事業經營法務碩士在職學位學程/碩士(2022年)

https://doi.org/10.6342/NTU202202116

摘要

無人載具，包含自動駕駛車輛及無人機等是已商品化的科技產品，並為各國及眾多廠商重要的發展方向，然而因為技術發展快速、影響層面不斷擴大及牽涉到生命安全，無人載具的資料法規研究有相當大的探索空間，特別有關載具科技、資料、個資及隱私間的研究。本文以個人資料保護角度探討自動駕駛車輛之資料運用。
自駕車被市場廣泛看好，並且已有多款商品化產品。自駕車集先進軟硬體技術於一身，被視為一台「移動的超級電腦」，近年來在快速成熟人工智慧技術推波助瀾下，更是如虎添翼，百花齊放的先進技術再再挑戰現行法規。自駕車資料法規研究的重要性在於：(一).自駕車深入生活，影響深遠;(二).巨量資料且更加以聯網車輛間資料之結合;(三).隱密不易被察覺;(四).事關安全，管理須謹慎;(五).人工智慧技術採用之必要;(六).法律上之「蒐集」、「處理」到「利用」行為難切割;(七).人工智慧模型資料須更新，監管困難;(八).人工智慧運用下資料與模型在地化;(九).少有對車外的資料或個資之探討;(十).自駕車目前以企業為主導，若為追求商業利益而運用個人資料者，將與個人權益有所衝突;(十一).國際已有自駕車個資之規範，相關法制亦應與時俱進，才能行穩致遠。
有鑒於自駕車發展是全球性的競爭，我國雖已於2018年通過《無人載具科技創新實驗條例》，目前台灣自駕車開發已進入測試階段，然而台灣現行《個資法》之相關規範是否足以涵蓋現有全球自駕車技術之快速發展與個資保護之趨勢?本文最後從企業責任與自我要求角度，以設計隱私（privacy by design）的概念應用到自駕車開發商或營運商應該要有的企業內部管理與企業責任。本文擬透過相關文獻先行定義出自駕車資料涵蓋之個資議題，其後透過分析與各國法規之比較，針對自駕車資料及個資議題提出管制之利與弊整理分析，以期未來在我國自駕車資料蒐集、處理及應用等議題之規範能夠貢獻出綿薄之力，並期待能夠提出實質有效的結論與建議。

關鍵字

人工智慧 ； 機械學習 ； 自駕車 ； 個人資料 ； 個資法 ； 隱私

並列摘要

Nowadays, unmanned vehicles, such as Automated Vehicles (AVs) and drones, are commercialized technological products, which are important development directions in various countries and for many manufacturers. However, the rapid development of related technology and its expanded influence can threaten personal safety and privacy. Considerable research can be done on the data regulations of unmanned vehicles, especially regarding personal data and privacy. This study discusses the data usage for AVs from the personal data protection perspective.
Automated Vehicle has a bright future in the market. Many major manufacturers and countries regard the development of AVs as an important direction of development, and many commercialized products already exist. Automated vehicles integrate advanced software and hardware and can be viewed as a “mobile supercomputer,” and such flourishing technology can challenge current laws and regulations. Self-driving technology can be divided into six levels, from zero to five, according to its degree of automation, ranging from the ubiquitous collision warning system to fully automatic driving similar to scenes depicted in the science fiction series Knight Rider. Behind all these self-driving technologies are a huge amount of raw data and calculations. Research on AVs data regulations should be done for several reasons: 1) Automated vehicles may become an integrated part of our lives and cause a far-reaching impact; by 2025, AVs in the automotive market are predicted to reach a penetration rate of 13%, and will continue to rapidly increase each year. Automated vehicles will roam the streets collecting and processing huge amounts of data, and will impact the country, society, and people; we have to be cautious when discussing the impact. 2) The sheer amount of data. 3) Data gathering can be secretive and difficult to detect. 4) It is a matter of safety. 5) Reliance on artificial intelligence. 6) Laws do not differentiate well enough between capturing, processing, and utilizing data. 7) Data need to be updated and is hard to regulate. 8) The localization of data and the models. 9) only a few discussions on the personal data collected outside of the car. 10) The development of AVs is mainly led by companies, and they could have different ideas that could cause disagreements on how personal data should be used. 11) There are relevant international AVs Data Protection regulation.
The development of AVs is a global competition. In 2018, Taiwan passed the “Unmanned Vehicles Technology Innovative Experimentation Act”, and is currently in its testing phase for AVs. However, although the personal data protection act has been revised twice in 2010 and 2015, are the regulations enough to cover the rapid global development of AVs? Currently, Taiwanese laws on AVs are made based on predictions of what the technology may look like in the future. Is it necessary to do a rolling review of the development of self-driving technology and change the regulations accordingly? In addition, from the perspective of corporate responsibility and self-discipline, the concept of “privacy by design” can be applied to the companies that develop and operate AVs. This study first identifies the problem surrounding the discussion on data collected by AVs in the relevant literature, then analyzes and compares the relevant discussions, guidelines, regulations, and drafts of various countries on AVs data and personal data. We analyze the pros and cons in the hopes that this study can provide a helpful conclusion and effective suggestions in the future regarding how AVs can collect, processing and use data in our country.

並列關鍵字

Artificial Intelligence ； Machine Learning ； Automated vehicles ； personal data ； personal data protection act ； privacy

參考文獻

期刊

1. WARREN, S. (1890). The Right To Privacy. Harv. L. Rev, 14(5), 193-220.

2. McCulloch, W.S., Pitts, W. (1943). A logical calculus of the ideas immanent in nervous activity. Bulletin of Mathematical Biophysics 5, 115–133. https://doi.org/10.1007/BF02478259

3. Turing, A. M. (1950). Computing machinery and intelligence. Mind, LIX, 433–460. doi:10.1093/mind/LIX.236.433.

4. William L. Prosser (1960). Privacy. California Law Review, Vol. 48, No. 3, 383-423.