A supervised system for curating browsing whitelists for individuals with cognitive disabilities under legal guardianship

Abstract

Ensuring digital safety for individuals under legal guardianship due to cognitive disabilities is the responsibility of their legally appointed guardians. These guardians are charged with making personal and lifestyle decisions in the best interests of the wards, which include their online access. Traditional blacklists are often ineffective for this purpose, as wards may trust misleading or harmful content across the entire web. Whitelists are a more suitable solution, as they restrict access to a predefined set of trusted websites, thereby reducing exposure to misleading, manipulative, or harmful content. However, creating and maintaining large whitelists is challenging and requires frequent updates, particularly for hyperlocal websites that are most important for a ward's social life and independence. We present SafeWeb, a supervised system designed to assist guardians in curating whitelists. Our method automatically curates relevant hyperlocal websites, which are essential for fostering ward independence but difficult to process with existing approaches. The processed websites are categorized into three groups: facility (e.g., schools and social centers), commerce (e.g., shops and service points), and prohibited (e.g., gambling venues and nightclubs). We provide the SafeWeb source code and curated websites for 20 countries (including USA, France, Japan, and Australia), along with validation results for selected countries. The results demonstrate SafeWeb's effectiveness as a decision-support tool for guardians to act in the ward's best interest.