OpenStreetMap Data Sourcing and Copyright Restrictions AI-pedia

OpenStreetMap (OSM) is a collaborative, open-access geographic database founded in 2004 by British entrepreneur Steve Coast. The project provides free geographic data to individuals, researchers, humanitarian groups, and commercial enterprises. Unlike proprietary mapping services such as Google Maps, TomTom, or Apple Maps, which restrict how their geographic data can be copied or redistributed, OpenStreetMap releases all of its data under an open legal license. Anyone may download, edit, distribute, and build commercial products from the database without paying licensing fees.

Because OpenStreetMap distributes its data under an open license, the legal integrity of its inputs is critical. Every piece of geographic information (referred to as geodata, which means data referencing specific geographic locations on Earth) must be legally clear of third-party intellectual property claims. If proprietary data is copied into the project, the entire database risks legal challenge from copyright owners. Consequently, OpenStreetMap maintains some of the strictest sourcing and copyright policies of any collaborative digital project.

Contributors are prohibited from tracing, copying, or deriving information from commercial digital maps, paper atlases, copyrighted tracking platforms, or unauthorized aerial and street-level imagery. Mapping within OpenStreetMap relies exclusively on primary data collected by contributors, explicit public domain resources, or datasets where rights holders have signed formal legal waivers. The community enforces these rules through technical oversight, volunteer auditing teams, and software tools designed to detect and remove infringing additions.

Legal Frameworks and Intellectual Property in Geodata

The rules governing data sourcing in OpenStreetMap emerge from three distinct areas of law: copyright law, database rights, and contract law (specifically website terms of service). Misunderstandings often arise because these legal concepts operate differently depending on the country where the contributor or rights holder resides.

The Idea-Expression Dichotomy and Factual Data

In standard copyright law, there is a fundamental distinction known as the idea-expression dichotomy. Copyright protects creative, original expressions of ideas, but it does not protect raw facts, concepts, or physical realities. For example, the physical location of a mountain peak, the coordinates of a road intersection, or the width of a river are objective facts about the natural and built world. In isolation, no individual or corporation can own the fact that a specific road exists at a specific latitude and longitude.

However, the way those facts are selected, coordinated, and visually presented can receive copyright protection. In the United States, the landmark Supreme Court case Feist Publications, Inc. v. Rural Telephone Service Co. (1991) established that compilations of facts are only copyrightable if they possess at least a creative "spark of originality" in their selection or arrangement. The court rejected the sweat of the brow doctrine, which had previously granted copyright protection based solely on the amount of labor, time, and money expended to compile factual information. Under United States law, raw coordinates are generally not copyrightable, but artistic cartography, symbology, and custom styling are protected.

Database Rights

In Europe, the legal reality is different due to the creation of database rights (specifically known as sui generis database rights, a Latin term meaning "of its own kind"). The European Union enacted Directive 96/9/EC on the legal protection of databases in 1996. This directive created a distinct form of legal protection that specifically rewards the "sweat of the brow" that United States copyright law rejects.

Under the European Union directive, if an organization makes a substantial investment in obtaining, verifying, or presenting contents, it can prevent the unauthorized extraction or re-utilization of the whole or a substantial part of that database. Because the OpenStreetMap Foundation is a company limited by guarantee registered in England and Wales, it is historically and legally bound by European and United Kingdom database protection laws. Therefore, even if a geographic dataset consists of pure, unadorned facts, extracting those facts from a protected European database without a license is a civil violation.

Contract Law and Terms of Service

Beyond copyright and database statutes, digital mapping platforms protect their assets through contract law. When an individual creates an account, browses a mapping website, or accesses an application programming interface (API, a software intermediary that allows two applications to talk to each other), they agree to a contract known as the Terms of Service (ToS).

Proprietary mapping providers routinely include contractual clauses that forbid users from reverse engineering, extracting data, or using their displays to create competing products. Even if an individual were in a legal jurisdiction where raw coordinates held no copyright or database protection, viewing a proprietary platform to capture those coordinates would constitute a breach of contract. OpenStreetMap rejects any data acquired through actions that violate the source's terms of service.

Licensing History and the Open Database License (ODbL)

OpenStreetMap's approach to sourcing is directly shaped by its license. The project has not always used its current legal framework; its evolution was prompted by the need to resolve ambiguities between artistic copyright and database law.

The Creative Commons Era (2004–2012)

At its inception, OpenStreetMap licensed its cartography and underlying vector data under the Creative Commons Attribution-ShareAlike 2.0 license (CC BY-SA 2.0). Creative Commons licenses were designed primarily for creative and artistic works, such as photographs, music, literature, and illustrations.

As the project grew, legal experts and the OpenStreetMap Foundation realized that CC BY-SA 2.0 had severe flaws when applied to geographic databases:

• Uncertainty over "works": It was unclear whether raw vector coordinates constituted an artistic "work" or a collection of facts exempt from Creative Commons terms.
• The ShareAlike trigger: CC BY-SA requires anyone who creates an "adaptation" or "derivative work" to publish the entire derived work under the same license. In cartography, it was legally ambiguous whether a business that placed proprietary delivery routes on top of an OpenStreetMap base map would be forced to make all of its private customer data public.
• Database right deficiencies: CC BY-SA 2.0 did not explicitly address European sui generis database rights, creating legal exposure across the European Union.

The Transition to ODbL

To establish legal certainty, the OpenStreetMap Foundation worked with legal specialists to develop the Open Database License 1.0 (ODbL), published by the Open Data Commons. The ODbL was built specifically for factual databases. It clearly separates the database right from the visual display and contents:

License Component	Scope of Coverage	Legal Application
ODbL 1.0	The Database Structure and Rights	Applies to the collection, ordering, and extraction of geographic records.
DbCL 1.0	Database Contents License	Waives individual copyright claims over the factual contents within the database.
CC BY-SA 2.0	Rendered Cartography (Historical/Tiles)	Historically used for standard visual map images (tiles) generated by the main server.

Under the ODbL, users may extract, modify, and build commercial operations using OpenStreetMap data. However, if a user modifies the OpenStreetMap data or merges it with other geodata to produce an improved geographic database (a Derivative Database), that user must publish their improvements under the ODbL as well. This reciprocal requirement is called a share-alike obligation.

The 2012 Redaction Process

Switching from CC BY-SA 2.0 to the ODbL required every contributing mapper to agree to the new terms. The OpenStreetMap Foundation initiated a relicensing campaign in 2010. While the vast majority of active contributors accepted the ODbL, thousands of historical contributors could not be reached, and some explicitly refused to sign.

On September 12, 2012, the OpenStreetMap Foundation executed the License Change Redaction. All data created or edited by non-agreeing contributors had to be stripped from the database to ensure the newly licensed dataset was legally pure. Over 100 million geographic points (known as nodes) and millions of lines (known as ways) were removed. In many regions, complete towns, roads, and coastlines disappeared overnight. The global community spent years running "re-mapping" campaigns, using permitted aerial imagery and ground surveys to rebuild the missing data. This experience proved to the community that accepting questionable data creates catastrophic long-term technical debt.

Prohibited Data Sources

To avoid another database contamination event, OpenStreetMap strictly categorizes which sources can and cannot be used. The fundamental rule is simple: if an external source has copyright restrictions, database rights, or terms of service that prevent open commercial reuse and re-licensing under the ODbL, it is strictly forbidden.

Proprietary Digital Mapping Platforms

Commercial web mapping services are the most common source of illicit copying. The community strictly forbids looking at, tracing from, or querying data from:

• Google Maps and Google Street View: Using Google Street View to verify a speed limit, check a street name, determine whether a path is paved, or trace building footprints is an infringement. Doing so breaches the Google Maps/Google Earth Additional Terms of Service and introduces proprietary information into OpenStreetMap.
• Apple Maps and Look Around: Apple’s map data and street-level imagery are proprietary and cannot be used as a reference.
• Here Technologies, TomTom, and Baidu Maps: Commercial datasets produced by traditional mapping firms are protected by copyright and database rights.

Commercial Paper Maps and "Trap Streets"

Traditional paper maps, guidebooks, and street atlases cannot be scanned, traced, or transcribed into OpenStreetMap. Commercial cartographers historically protected their work using trap streets (also called paper towns or fictitious entries). A trap street is a deliberately invented, nonexistent street, alleyway, or typographical error inserted into a map by its creator.

If a competitor copies the map verbatim, the trap street will appear in their product. Because the feature does not exist in the physical world, the original cartographer can present this in court as irrefutable evidence of copying rather than independent physical surveying. OpenStreetMap mappers who copy from commercial paper maps risk copying these trap streets, immediately exposing the OpenStreetMap Foundation to copyright lawsuits.

User-Generated Platforms and Third-Party Fitness Apps

A widespread point of confusion among new contributors involves user-generated content (UGC) websites and outdoor recreation platforms, such as Wikiloc, AllTrails, Strava, and Komoot. A common question among outdoor enthusiasts illustrates this exact dilemma:

"If someone sees a trail on Google Street View and has multiple independent Wikiloc tracks using it, does that make it permissible to add to OpenStreetMap, or is it a copyright violation?"

Adding this trail to OpenStreetMap is a violation on two distinct grounds:

1. Google Street View is off-limits: Using Google Street View to locate, confirm, or verify the geometry or existence of a trail derives data from Google’s proprietary system. This breaches Google's terms and taints the OpenStreetMap database.
2. Wikiloc tracks remain individual intellectual property: Wikiloc is a platform where individuals upload GPS tracks. Under the terms of service of Wikiloc and international copyright frameworks, the copyright of those recordings belongs to the individual uploaders, while Wikiloc retains platform distribution rights. Wikiloc’s general website terms do not license those tracks for third-party extraction or commercial redistribution under the ODbL. Even if five different people have walked a path and posted their tracks to Wikiloc, a third party has no legal right to trace those tracks into OpenStreetMap without explicit permission from every single track owner.

The assumption that multiple independent users uploading a path transforms private data into "public domain facts" is a legal fallacy. Combining two unauthorized sources (Google Street View and Wikiloc) does not create a lawful derivative; it simply compounds two separate legal violations into a single OpenStreetMap edit.

Unapproved Government and Institutional Portals

Not all government data is free for use. While many national mapping agencies make their geodata publicly visible on the web, visibility is not equivalent to an open license. For example, data from the Ordnance Survey in the United Kingdom or the Institut national de l'information géographique et forestière (IGN) in France historically carried strict crown or state copyright. Contributors cannot use institutional sources unless the institution has explicitly issued an open license that is formally compatible with the ODbL.

Permitted and Compatible Data Sources

OpenStreetMap relies on three primary categories of permissible data: direct personal surveys, authorized imagery, and compatible open government datasets.

Ground Surveying and Direct Field Observation

The original, most legally unassailable source of data is direct field observation. When a contributor walks, cycles, or drives along a physical road, they are making primary observations of physical reality. No third party holds a copyright over the physical presence of a fire hydrant, a storefront name, a traffic sign, or a path across a public common.

Survey methods include:

• Personal GPS Logs: Mappers record their own journeys using a global positioning system (GPS) receiver or smartphone, generating files in the GPX format (GPS Exchange Format, an open XML standard for coordinates). These raw recordings belong entirely to the mapper, who freely submits them to OpenStreetMap.
• Field Papers: A process where mappers print out a paper map of a neighborhood, walk through the area writing notes and drawing amendments by hand, and then manually transcribe their personal observations into an editor.
• Survey Software: Mobile apps like StreetComplete, Every Door, and Vespucci allow mappers to walk through an environment and answer factual survey questions (such as "What is the speed limit here?" or "Does this shop have wheelchair access?") directly into the database.

Authorized Aerial and Orthoimagery

Tracing roads, buildings, waterways, and land use from satellite and aerial photography is the most common way vector data is created in modern OpenStreetMap. However, satellite photography is expensive to produce and enjoys strong copyright protection. Tracing from aerial imagery produces a derived work; therefore, mappers may only trace from imagery providers who have granted explicit legal permission to OpenStreetMap.

The standard editing software (such as the web-based iD editor or the desktop-based JOSM, the Java OpenStreetMap Editor) comes pre-configured with authorized imagery layers. Major permitted providers include:

• Bing Maps: In November 2010, Microsoft granted OpenStreetMap the legal right to trace vector data from its global Bing aerial photography layer. This was a turning point for the project, enabling global tracing at high resolution.
• Esri World Imagery: Environmental Systems Research Institute (Esri) provides an authorized aerial layer to OpenStreetMap contributors to support map accuracy.
• Mapbox Satellite: Mapbox provides high-resolution imagery specifically cleared for OpenStreetMap tracing.
• Maxar Technologies: Maxar has periodically provided high-resolution imagery layers to OpenStreetMap, particularly for disaster-relief mapping overseen by the Humanitarian OpenStreetMap Team (HOT).
• State and Municipal Orthoimagery: Many regional governments provide orthoimagery (aerial photographs that have been geometrically corrected so that the scale is uniform and matches a map projection). Mappers may use these if the local government has issued an ODbL-compatible license.

Authorized Street-Level Imagery

While Google Street View and Apple Look Around are strictly banned, there are street-level photographic platforms explicitly designed to be compatible with OpenStreetMap:

• Mapillary: Mapillary is a crowd-sourced street-level imagery platform owned by Meta. Mapillary's terms of service include a specific, legally binding exemption that grants OpenStreetMap contributors the right to use its images and derived metadata to improve the OpenStreetMap database.
• KartaView: Formerly known as OpenStreetCam and originally developed by TeleNav, KartaView is an open, crowdsourced street-level photograph platform that provides permissive access for OpenStreetMap editing.

Open Government Datasets and Public Domain Works

In certain countries, federal legal doctrines place government works directly in the public domain. In the United States, works created by federal government employees as part of their official duties are not eligible for copyright protection under Title 17, Section 105 of the United States Code.

Consequently, early OpenStreetMap developers were able to import the United States Census Bureau's TIGER dataset (Topologically Integrated Geographic Encoding and Referencing). While TIGER had significant spatial errors and required extensive cleanup, the import was completely lawful. Similarly, data produced by the United States Geological Survey (USGS) and National Oceanic and Atmospheric Administration (NOAA) can be safely utilized.

In other nations, specific open data licenses have been created to allow reuse. When a government releases data under a license such as the UK Open Government Licence (OGL) or the French Licence Ouverte, the OpenStreetMap Foundation’s legal team examines the text to ensure it does not conflict with the ODbL’s share-alike rules. If compatible, mappers may incorporate those public records.

Concepts of Derivation and "Tainting"

Understanding copyright in OpenStreetMap requires an understanding of what constitutes a derivative work and how intellectual property violations create systemic risks for the entire project.

What Constitutes Derivation?

A derivative work is a creation based upon or derived from one or more preexisting works. In geographic data, derivation does not only mean copying a raw file. It includes any workflow where proprietary information informs the content of the new vector entry:

• Tracing: Manually following the lines of an aerial photograph, vector map, or scanned image to produce a sequence of coordinates.
• Transcription: Looking at a proprietary map to read text, such as street names, house numbers, or bridge height limits, and typing those details into OpenStreetMap.
• Verification ("Eyeballing"): Looking at a commercial map to confirm whether a road connects to a highway, or whether a trail actually exists. Even if the mapper draws the path using their own rough memory, verifying the topology against a prohibited source constitutes deriving data from that source.
• Offset correction: Adjusting existing OpenStreetMap data to match the layout of an unapproved proprietary map.

The "Taint" Phenomenon

When an unauthorized derivative work is committed to OpenStreetMap, the database suffers from what developers and licensing experts term taint. Geographic data behaves like a connected graph: nodes form ways, ways connect at intersections, and intersecting ways establish complex routing and administrative relationships.

If an editor injects an unauthorized trail or road network into a national park, downstream consumers—such as logistics companies, emergency dispatchers, search engines, and smartphone app developers—download that data. If the true copyright holder takes legal action, the presence of the tainted data can:

1. Expose commercial consumers of OpenStreetMap to copyright infringement claims.
2. Force downstream platforms to purge their local caches and spatial indexes.
3. Compel the OpenStreetMap Foundation to carry out costly structural redactions.

For this reason, the community treats even minor copyright violations not as isolated occurrences, but as a form of database corruption that threatens the legal viability of the entire platform.

Detection, Enforcement, and the Data Working Group

Because OpenStreetMap is open for public editing, malicious or misinformed edits happen daily. To protect the database from legal liability, the project relies on community oversight and a specialized administrative body.

The Data Working Group (DWG)

The Data Working Group (DWG) is a volunteer committee appointed by the OpenStreetMap Foundation Board of Directors. The DWG has legal authority to investigate copyright disputes, resolve vandalism, enforce community policies, and interact with the Foundation's legal counsel.

When a copyright infringement is reported, the DWG undertakes an investigative workflow:

1. Interviewing the contributor: The DWG contacts the mapper responsible for the edit via the OpenStreetMap messaging system, asking them to identify the source of the questionable data.
2. Examining edit patterns: Investigators review the user's historical changesets (a changeset is a group of edits committed at one time by a user, which includes a comment describing the source). They look for suspicious metadata, such as changeset comments saying "copied from Google" or geometries that mirror proprietary commercial road curves down to sub-millimeter precision.
3. Issue of blocks: If a user fails to respond, gives evasive answers, or persists in using banned sources, the DWG places a block on the user’s account. Blocks can be temporary (requiring the user to log in and read an administrative message) or permanent.

Software Detection Tools

The community monitors additions using automated auditing software. Tools such as OSMCha (OpenStreetMap Changeset Analyzer) allow monitors to flag edits that contain suspicious keywords in their source tags (for example: "google", "waze", "here", or "wikiloc").

Advanced algorithmic tools compare vector alignments. If an editor uploads a massive road network in an area where permitted aerial imagery is obscured by clouds, monitors investigate how the user could have identified those paths without an unauthorized reference source.

The Mechanical Redaction Tool

When the DWG confirms that infringing data has entered the system, standard user deletion is insufficient. OpenStreetMap maintains a complete, permanent historical record of every change ever made to an element. If an infringing road is simply "deleted" by a normal mapper, the copyrighted coordinates still remain in the historical API database and can be accessed via version history.

To eliminate legal liability, the DWG uses a specialized server-side tool called the Redaction Bot. The redaction software completely purges the specific infringing revisions from the public database and API view. The history tables are modified so that the proprietary data is hidden from downstream users, restoring the database to a legally defensible state.

Edge Cases, Ambiguities, and Ongoing Debates

Rapid technological changes regularly create new challenges for OpenStreetMap's copyright standards. Several areas currently present complex legal and technical questions.

Artificial Intelligence and Machine Learning Vectorization

A major development in modern cartography is the use of machine learning algorithms to automatically detect features in high-resolution satellite imagery. Companies like Meta (through the RapiD platform and Daylight Map distribution) and Microsoft produce massive datasets of computer-predicted building footprints and road centerlines.

The community has engaged in extended debate over the use of these tools:

• Source imagery rights: If an artificial intelligence model is trained on proprietary satellite imagery that OpenStreetMap does not have permission to trace, does the resulting vector output constitute an unauthorized derivative work of the imagery?
• Model weights and copyright: Intellectual property law has not yet resolved whether machine learning models retain the copyright status of their training inputs.
• Community consensus: OpenStreetMap currently permits the use of AI-derived features (such as Microsoft Building Footprints) only if the provider explicitly licenses the output under ODbL-compatible terms and the imagery used to run the inference was legally cleared for tracing.

Aggregated GPS "Heatmaps"

While tracing an individual's private GPS recording without their permission is a copyright violation, many athletic platforms generate aggregated heatmaps. A heatmap is a visual representation composed of thousands of overlapping, anonymized GPS tracks recorded by millions of users. The Strava Heatmap is the best-known example.

In the past, Strava granted OpenStreetMap permissions to trace from its global slide and heatmap imagery to correct path geometries. However, licensing terms for commercial heatmaps fluctuate frequently. If a company revokes access or alters its terms of service, mappers must immediately stop using the layer. Furthermore, because a heatmap is an aggregated visual product, debate continues over whether tracing from it extracts protected user content or merely reflects factual public path usage.

Public Domain and Freedom of Panorama

Another subtle edge case involves photography and street-level mapping. Freedom of panorama is a legal provision in copyright law that allows people to take photographs of buildings, sculptures, and artwork located in public spaces without infringing the architect’s or artist’s copyright.

In countries with weak or non-existent freedom of panorama (such as Italy or South Korea), taking photographs of public buildings and using automated computer vision to extract their facades into an open database can create legal complications. While OpenStreetMap primarily stores vector outlines rather than photographic textures, mappers documenting artistic installations, historic monuments, or structural details must occasionally navigate these jurisdictional discrepancies.

Best Practices for Contributors

To maintain the legal security of OpenStreetMap, mappers must observe systematic habits during data collection and editing. Following established guidelines prevents accidental contamination.

The Contributor Verification Checklist

Before adding any piece of geographic data, a mapper should be able to answer "yes" to at least one of the following questions:

• Did I personally visit this physical location and observe this feature with my own eyes?
• Did I personally record the GPX track or take the photograph being used as a reference?
• Is this feature traced from an imagery layer that is formally verified and pre-approved within an official editor (such as standard Bing, Esri, or Mapbox layers)?
• Is this data imported from an open government source that has been evaluated, approved, and documented on the OpenStreetMap Community Wiki?

If the answer to all of the above is "no," the data cannot be added.

Proper Attribution and Source Tagging

OpenStreetMap utilizes a tag-based metadata system. Every object consists of geographic coordinates accompanied by key-value pairs (e.g., highway=footway). Whenever data is created or edited, contributors are expected to populate the source=* tag on either the individual object or the overarching changeset.

Source Type	Recommended Changeset Tag Example	Legality
Field Survey	source=survey or source=local_knowledge	Permitted; highest standard of legal safety.
Personal GPS Recording	source=gpx	Permitted; must be the mapper's personal track.
Approved Aerial Imagery	source=Bing or source=Esri World Imagery	Permitted; imagery must be authorized.
Approved Street-Level Views	source=Mapillary or source=KartaView	Permitted; covered by explicit provider waivers.
Commercial Web Maps	source=Google Maps or source=Street View	Strictly Prohibited; results in immediate reversal and account block.
Third-Party Trailing Apps	source=Wikiloc or source=AllTrails	Strictly Prohibited; unapproved extraction of private copyright.

Accurate source documentation is essential. If a monitor reviews a changeset that modifies a complex network of forest trails but leaves the source tag blank, that changeset is immediately suspect. By clearly identifying the permitted imagery, personal track, or authorized open dataset used, contributors save the community hundreds of hours of auditing labor and ensure that OpenStreetMap remains a dependable, legally resilient public utility.

See also

• Creative Commons
• Database Directive (European Union)
• Feist Publications, Inc. v. Rural Telephone Service Co.
• Geographic information system (GIS)
• Mapillary
• Open Data Commons
• Open Database License (ODbL)
• OpenStreetMap Foundation
• Sweat of the brow
• Trap street