Look Around in Apple Maps on an iPhone
Apple has hundreds of terabytes of data for its Apple Maps Look Around feature, and the company is using that data to train generative AI models.
Look Around is a function of Apple Maps that gives users a road-level viewpoint of an area, similar to Google Street View. However, the data collection effort fueling the feature will also be used for AI purposes.
In an updated declaration about Apple Maps Image Collection spotted by 9to5Mac, Apple confirms it will be using any data collected in the process to improve its AI models.
“In addition to improving Apple Maps and the algorithms that blur faces and license plates in images published in Look Around feature, Apple also will use blurred imagery collected during surveys conducted beginning in March 2025 to develop and improve other Apple products and services,” reads part of the declaration.
“This includes using data to train models powering Apple products and services, including models related to image recognition, creation, and enhancement.”
Taking the description at face value, this would mean features like Visual Intelligence, Photo’s Clean Up feature, and Image Playground as part of Apple Intelligence, among others. It could also be employed to train other models for yet-to-launch models that Apple is still working on.
A trove of environmental data
While normal mapping data is useful in general, the information gathered for Look Around offers a lot more elements for AI model training and for other areas.
The data is collected via a specially prepared backpack worn by a pedestrian Apple employee or contractor, or on a rig attached to a dedicated vehicle. Depending on the rig, this can include capturing not only images of an area, but also 3D depth mapping data.
Evidently, for privacy reasons, Apple will be using the final imagery used by Look Around for its training, since it will have blurred people’s faces and license plates before publication.
The move to use the data is to be expected, as training AI models can require lots of information, and the more information, the better. This has led to controversy over the sourcing of data, including paid sources and scraped information.
Some companies are taking somewhat unethical steps to get the data, such as Google’s argument that copyright laws shouldn’t apply to AI training, as well as Adobe’s terms debacle.
On Apple’s side, it is at least trying to be ethical, paying for access to data instead of simply scraping it. Reusing data collected for Apple Maps certainly makes sense for Apple to do, since it collected the data for itself and doesn’t have to ask permission to use it in other ways.
