The Enshittification of Google Maps
Hi friends,
Almost every time I drive down the West Side Highway in New York City, Google Maps tells me to get off at the ship terminal – even though I am not aiming to set sail.
The algorithm seems to be attracted to the strip of road parallel to the highway that has fewer cars on it – even though it is a ship passenger loading and unloading zone. Technically, the app is likely correct that it would be faster to use the loading zone to zip ahead of highway traffic and then merge back onto the highway - but it feels morally wrong and anti-social, the same way that it feels wrong to use the highway shoulder to zip around traffic.
This is the problem of the Google Maps algorithms that we all increasingly rely on. Google, with its 80 percent market share in maps – is optimized to find the shortest route, even just a minute shorter, often at the expense of other important factors like considering what a road is designed to be used for. In my latest New York Times Opinion piece (gift link), I argue that Google’s monopoly on maps has slowed innovation that might have promoted better routing options.
Google Maps only allows users the most rudimentary controls: “avoid tolls” and “avoid highways.” In today’s world, having so little control of our routes feels so old-fashioned that it is practically medieval. On social networks, I can usually block users, mute keywords, and even hide ads that I don’t like. And even then, I want more control of my social media feeds so that I can ensure I am seeing posts by my friends and trusted influencers.
Writer Cory Doctorow has a term for Big Tech’s declining user focus: enshittification. It’s when a monopoly realizes it doesn’t need to spend time improving quality for the users because they have so much market share. But it’s not a term allowed in the New York Times ;-)
The real life consequences of enshittification can be devastating. Communities have suffered from the apps almighty focus on shaving off a few minutes of time. Jurisdictions ranging from Los Angeles to the beach town of South Shores, North Carolina have begged Google to stop diverting high volumes of traffic from highways onto residential streets. So many trucks have gotten stuck on a winding Vermont road called Smugglers’ Notch that lawmakers have introduced legislation to fine navigation apps if they don’t warn that trucks are prohibited on the road.
In 2022, Philip Paxson drowned on his way home from his daughter’s 9-year-old birthday party after Google maps directed him to drive off a collapsed roadway that fed into a washed-out bridge. His car plunged into a creek. His family sued Google for negligence. According to the family’s lawsuit, a local resident had submitted an edit to Google maps two years earlier noting the washed-out bridge and asking for the map to be updated. But at the time of the accident, it was still not updated.
Google’s monopoly extends to its geographic data, too. Google owns its own proprietary map data, generated in part by the Street View cars that it fields across the globe. Almost every other map relies on OpenStreetMap data, collected and vetted by volunteers, which is free to use (as long as you credit the source).
In this open data set lies an opportunity for a future with more navigational choices. But OpenStreetMap, sadly, has nowhere near the staff and resources needed to support consumer use. Its maps are remarkably high quality and used as a base layer by all the leading players including Amazon, Facebook and Apple. But OpenStreetMaps global budget is shockingly tiny – at just about $1 million a year. By comparison, the foundation that supports Wikipedia spent nearly $170 million in 2022.
OpenStreetMap is so under-resourced that it only recently hired paid systems administrators to ensure its backend machines kept running. It doesn’t have the staff to take in data that companies such as Lyft and Uber want to share with them, says Maggie Cawley, executive director of OpenStreetMap US. “It’s a very fragile ecosystem,” she told me.
So if we want a more robust and competitive public map infrastructure, we should start by supporting OpenStreetMap. The giant corporations that rely on its maps should stop freeloading and financially support the project. City planners should dedicate staffers to update their roads and bridges on OpenStreetMap. Communities should submit updates to OSM with roads and hazards in their local areas.
After all, if there is one thing we should have all learned from the AI revolution, it is that data is its essential ingredient. And so if we want to build better AI navigation, we need to protect and nurture that geographic data that will fuel that future.
As always, thanks for reading.
Best
Julia Angwin