On Privacy Disclosure from User-Generated Content of Automation Rules

Trigger-Action Platforms (TAPs) are systems that enable users to automate routine tasks, such as turning off lights at a specific time, without requiring technical skills. In the process of creating automation rules, users are prompted to provide descriptions in natural language, which are referred to as User-Generated Content (UGC), such as the title that explains the intended behavior of the rule. However, UGC may contain sensitive information that could expose users to unwanted situations or be exploited by cyber attackers. This position paper provides an initial assessment of the risks associated with UGC in TAPs and discusses the use of NLP techniques to mitigate these risks. Additionally, the paper highlights the need for further research to better understand the impact of UGC on privacy and to develop effective privacy-preserving mechanisms for TAPs.