Noisy
← this week  ·  Subscribe →

The world is being quietly rearranged by people who write very long documents.

arXiv · April 7, 2026
The title they went with InsTraj: Instructing Diffusion Models with Travel Intentions to Generate Real-world Trajectories Noisy translates that to

You can now generate fake GPS routes from English text — raising privacy and urban planning questions

What happened
Researchers built a system that turns natural language descriptions into realistic GPS trajectories using AI. This means urban planners and researchers can now generate synthetic movement data without exposing real people's actual routes.
Why it matters
Until now, if you wanted realistic GPS data to test urban planning ideas or train autonomous systems, you had to use actual human location data — or use obviously fake, unrealistic synthetic data. This system splits the difference: it can generate movement patterns that look human and follow specific constraints (avoid highways, stay in downtown, move like a delivery driver) without recording anyone's actual location. The catch is obvious: the better synthetic GPS data gets, the easier it becomes to generate convincing fake movement patterns for surveillance, stalking, or deception without any real person having to be followed.
The signal
Watch whether urban planning departments or ride-sharing companies adopt this for testing. The adoption pattern will tell you whether the synthetic data quality is actually good enough for real-world use, or whether researchers are still treating it as a lab curiosity.
If you insist
Read the original →