Location sharing is creating a massive dataset based on your personal movements.
Washington Post illustration.
Nov. 14, 2024 — In today’s Washington Post, writer Tatum Hunter explores an issue that’s quietly emerging in the AI world.
Her article, “Location sharing is making us miserable. It’s time to say ‘no,’” dives into the ways that our smartphones keep sharing our geolocation data with friends and family via a variety of apps.
She writes:
Location sharing is the privacy problem that crept into our lives when we were too busy to notice its costs. While we work multiple jobs, juggle family commitments and drive for hours every day, location sharing helps hold things together. People increasingly rely on apps like Find My Friends, Snapchat or Life360 to keep track of their inner circles — or even acquaintances. What’s the harm of a little surveillance among friends?
Hunter adds: “But you wouldn’t hand over access to your text messages or Google search history. Location sharing puts the same sort of intimate information — your daily movements — on display.”
Her Post article focuses on the interpersonal risks involved in all this location sharing. A demanding friend or over-involved parent may use geolocation data to cross the boundary separating fun from intrusive. More seriously, domestic abusers may use geolocation data to track, control, or harm their victims.
The ways in which our smartphones, and their apps, constantly track an individual’s movements throughout the day isn’t just an interpersonal issue, however.
In the world of AI and data privacy, there’s growing concern about the collection and use of personal data in the massive training datasets used to create AI models. That personal data could include information about an individual’s movements based on geolocation data tracked by a variety of apps, everything from Google to Snapchat to AccuWeather to park-by-phone services.
It’s not farfetched to envision AI systems capable of compiling a detailed dossier on an individual based solely on their geolocation data: Where they shop for groceries, what medical offices they visit, what stores they patronize, where they get gas, whose homes they visit, where their kids go to school. The list is as endless as the places each of us go in the course of our daily lives.
new California law brings ‘personal information’ into AI age
California’s recently adopted AB 1008 aims to prevent the sharing of personal information by AI systems, by clarifying the definition of personal data under the existing California Consumer Privacy Act (CCPA).
The 2018 CCPA already included geolocation data in its definition of personal information. What AB 1008 does is specify that personal information can exist in various formats, including within artificial intelligence systems. In other words, personal information like geolocation data does not cease to be personal information (that is, protected under the CCPA) simply because it’s been washed through an AI system.
The Transparency Coalition supported AB 1008 as it made its way through the California legislature, and we are encouraging the introduction of similar laws in other states in 2025.