Skip to content
Search

Latest Stories

Top Stories

Government must protect us from geolocational disinformation

City with GPS markers
Mongkol Chuewong/Getty Images

Crampton is an adjunct professor of geography at George Washington University and a member of Scholars.org.

As artificial intelligence is becoming more powerful and is more embedded in society, global governments are beginning to regulate these types of technology, and to balance benefits with potential harms. Yet while significant attention has been paid to reducing risk in the realms of health, finance and privacy, policymakers have left one element largely unaddressed: geolocation data.


This data — which provides information on the physical location of a person or device like a smartphone — is powerful, sensitive and highly valuable. AI procedures that are already being adopted to acquire, process and automate spatial and locational data are a particular concern that call for swift action. But policymakers can simultaneously look to the future and work to ensure that we develop independent, trustworthy AI governance for geolocation by drawing on the hard-won knowledge of the spatial digital revolution of the past two decades. To realize the best outcomes on privacy, combatting disinformation and deanonymization threats, policymakers must partner with geospatial domain experts rather than legislate around them.

The current regulatory landscape

In 2023, privacy legislation protecting “sensitive information” was passed in California, Colorado, Connecticut, Delaware, Florida, Indiana, Iowa, Montana, Oregon, Tennessee, Texas and Utah. A significant number of these laws include a provision covering “precise geolocation.” Data that qualify as indicating precise geolocation are limited to a radius of 1,750 feet around their subject, the equivalent of approximately one-third of a mile — significant territory in a densely populated urban area, as this interactive map of health care facilities in Washington, D,C., shows.

In Europe, the EU AI Act has prohibited the real-time collection of biometric data, such as occurs in facial recognition technology. Concurrently, U.S. legislators have become increasingly concerned about the risks of artificial intelligence, and the White House issued an executive order calling for new standards to prevent AI bias, threats or misuse. In 2018 in Carpenter v. United States, the Supreme Court held that law enforcement agencies require a warrant to obtain location data from cell-phone towers; however, this data is imprecise, and law enforcement actors switched to using richer data sources not covered by the ruling (like app-based location data sourced directly from Google).

While these are welcome developments in the ongoing need to secure privacy, geolocation has certain unique risks that this legislation and the policymakers concerned with it have yet to address.

Risks

There are three main categories of risk for geolocation data governance: disinformation, surveillance and antitrust/concentration of market. Disinformation (e.g. fake maps and data, propaganda), bias, and discrimination raise issues of trustworthiness, privacy, and ethics. The concern for AI is not just low-quality knowledge, but low-quality learning and low-quality meaning. For example, predictive policing — where data is analyzed to predict and intervene on potential future crime —may be based on poor, false or biased data that can lead to real-world discriminatory consequences.

  • Fake data can infiltrate maps (spatial databases), intentionally or not. Fake data could be included in driving apps or autonomous vehicles to chaotic effect; AI could be applied to geospatial data in a manner that misclassifies satellite imagery.
  • Disinformation may include falsely showing a person to be at a location when they were not — known as “location spoofing” — for blackmail or to cause reputational damage. “ Deepfake geography ” involves faking that a person is not at a place they should be: Imagine truckers’ data hacked to falsely show them as having deviated from their routes.
  • Inadvertent misinformation proliferated by lack of relevant geospatial analytic expertise can lead to detrimental outcomes. Inaccuracy and uncertainty can arise from analyzing a phenomenon at the wrong spatial scale (known as the “ the Openshaw Effect ”), or not accounting for how boundaries influence the scale of the analysis of aggregated data (referred to as “modifiable areal unit problem”).

Surveillance and locational tracking, which can include wide-scale biometric identification in real time or upon review of previously gathered data, poses many threats to privacy. Inference of personally identifiable information based on geospatial data obtained through surveillance is all too easy, and can include privacy infringements like re-constituting encrypted data (deanonymization) and uncovering the identity of a person or organization that has been obscured (re-identification.)

One well-known 2013 study found that knowing just four location points was enough to re-identify 95 percent of individuals, and that even when geospatial data is less precise, it can still reliable re-identify individuals — it takes a high factor of imprecision before location data loses its power to pinpoint. A new study of metro card travel data confirmed the findings that three random location points from within a period between one minute and one hour are sufficient to identify 67 percent to 90 percent of users. And facial recognition technology, which has many clear privacy risks, is now widely employed by law enforcement.

The limits of antitrust regulation and market concentration among tech companies point to increased opportunities for large data breaches or unethical use. The market is dominated by deep-pocketed AI tech companies including OpenAI, Google and Microsoft; these companies own and control the high-tech market, especially “high compute” fields like machine learning and AI training, effectively locking out competitors.

Recommendation to begin risk mitigation

The Offices of Science and Technology Policy at the White House and Congress can hold hearings with geospatial industry and academic experts to identify current and emerging threats to privacy from geolocation data and geolocation services and analytics. The quality and efficacy of legislation will depend on collaboration with and transparency from the experts who are designing and deploying these emerging technologies.

Read More

Biased Coverage Distorts the Historical Record We Later Inherit
white printer paper on black table
Photo by Ashni on Unsplash

Biased Coverage Distorts the Historical Record We Later Inherit

I used to enjoy doing my schoolwork in my college newspaper’s office. There is a series of tall library shelves filled with dusty books held together by loose binding that contain every article printed since our inception in the 1930s.

The book covers have lost the sharpness of their hues over time, and the thin old papers inside are yellow and torn, but inside those books lie almost 100 years of articles that tell the stories and history of the college town, Isla Vista, and UC Santa Barbara, as written by student journalists at the Daily Nexus.

Keep ReadingShow less
Media criticism
News media's vital to democracy, Americans say; then a partisan divide yawns
Tero Vesalainen/Getty Images

Public Media Under Fire: Why Project 2025 Is Reshaping NPR and PBS

This past spring and summer, The Fulcrum published a 30-part, nonpartisan series examining Project 2025—a sweeping policy blueprint for a potential second Trump administration. Our analysis explored the proposed reforms and their far-reaching implications across government. Now, as the 2025 administration begins to take shape, it’s time to move from speculation to reality.

In this follow-up, we turn our focus to one of the most consequential—and quietly unfolding—chapters of that blueprint: Funding cuts from NPR and PBS.

Keep ReadingShow less
Medical Schools Are Falling Behind in the Age of Generative AI

"To prepare tomorrow’s doctors, medical school deans, elected officials, and health care regulators must invest in training that matches the pace and promise of this technology," writes Dr. Robert Pearl.

Getty Images, ArtistGNDphotography

Medical Schools Are Falling Behind in the Age of Generative AI

While colleges across the nation are adapting their curricula to harness the power of generative AI, U.S. medical schools remain dangerously behind.

Most students entering medicine today will graduate without ever being trained to use GenAI tools effectively. That must change. To prepare tomorrow’s doctors – and protect tomorrow’s patients – medical school deans, elected officials, and health care regulators must invest in training that matches the pace and promise of this technology.

Keep ReadingShow less
Bay Area Social Media Post Claims ICE Cannot Enter Library, Fuels Misinformation

South Novato Library, California

Pricila Flores

Bay Area Social Media Post Claims ICE Cannot Enter Library, Fuels Misinformation

Bay Area community advocates are cautioning community members to be wary of what they see, interact with, and post on social media regarding information about the United States Immigration and Customs Enforcement (ICE) and immigration, following a rumor that targeted the Marin County Library.

‘South Novato Library has safe rooms that cannot be accessed by border patrol or ICE without a court order,’ an Instagram story post reads, with photos of a room in the library next to the text alongside the library address. The graphic claims Immigration and Customs Enforcement would not have the right to enter the pictured room without a court-ordered warrant.

Despite the graphic becoming a popular share among the local community of Novato, a Marin County city located just north of San Francisco, the information is false.

Keep ReadingShow less