Personal autonomy is at the core of liberal societies, and its preservation has been a focus of European Union (EU) consumer and data protection law. Professionals increasingly use artificial intelligence in consumer markets to shape user preferences and influence their behaviours. This paper focuses on the long-term impact of artificial intelligence on consumer autonomy by studying three specific commercial practices: (1) dark patterns in user interfaces; (2) behavioural advertising; and (3) personalisation through recommender systems. It explores whether and to what extent EU regulation addresses the risks to consumer autonomy of using artificial intelligence in markets in the long term. It finds that new EU regulation does bring novelties to protect consumer autonomy in this context but fails to sufficiently consider the long-term consequences of autonomy capture by professionals. Finally, the paper makes several proposals to integrate the long-term risks affecting consumer autonomy in EU consumer and data protection regulation. It does so through an interdisciplinary approach, drawing from legal research and findings in the study of long-term thinking, philosophy and ethics and computer science.

个人自主权是自由社会的核心，其保护一直是欧盟 (EU) 消费者和数据保护法的重点。专业人士越来越多地在消费市场中使用人工智能来塑造用户偏好并影响他们的行为。本文通过研究三种具体的商业实践，重点关注人工智能对消费者自主权的长期影响：（1）用户界面中的深色模式；(2)行为广告；(3) 通过推荐系统实现个性化。它探讨了欧盟监管是否以及在多大程度上解决了长期在市场上使用人工智能给消费者自主带来的风险。报告发现，新的欧盟法规确实为保护消费者自主权带来了新颖性，但未能充分考虑专业人士夺取自主权的长期后果。最后，本文提出了几项建议，将影响消费者自主权的长期风险纳入欧盟消费者和数据保护监管。它通过跨学科的方法来实现这一目标，借鉴了法律研究以及长期思维、哲学和伦理学以及计算机科学的研究成果。