Facebook has been looking for contractors who can help it interpret mapping data to deliver internet access to remote parts of the world, according to job listings and people approached by the company.

The company has been seeking editors and quality analysts for its Maps team, whose jobs would focus on improving internet access in rural areas and finding areas that may need disaster relief. According to the job listings, the hires would analyze and check data from OpenStreetMap, a project that uses crowdsourced information to provide a map of the world.

One person who was approached by outside recruiters on behalf of Facebook about a potential job opportunity with these teams was told the project would have a “humanitarian” focus. One other current employee wrote on their LinkedIn profile that the team uses satellite imagery to help improve internet access. Many people have joined the team over the last year, according to LinkedIn profiles.

The company told CNBC it has been open about its work with OpenStreetMaps, which it is using to create detailed street maps of underprivileged areas where people are not represented on typical maps, and it has other humanitarian mapping projects, such as building maps of disaster areas.

Facebook has also used satellite imagery to map populations in more than 20 countries as it explores ways to bring internet access to remote areas.

Facebook had 1.45 billion daily active users as of March, and it has estimated that 10 percent of the world’s population lacks internet access.

Although these roles are for humanitarian projects, Facebook may someday have commercial uses for mapping experts. Last year, Facebook rival Snap launched Snap Maps, which shows where Snapchat users are posting their snaps. Snap Maps uses OpenStreetMap and Mapbox data, combined with satellite imagery from Digital Globe. Facebook has a recent history of copying popular features from Snapchat, particularly certain Instagram features.

In addition, Facebook filed two patents in August 2017 for a system that would use map data from multiple map sources and another that would automatically detect location, according to the Silicon Valley Business Journal.

– Additional reporting by Jordan Novet