Courts lock up public databases

AP Photo by Will Shilling

Ohio Supreme Court Justice Paul Pfeifer said that requesters were being forced to pay for a government’s decision to use systems that make retrieving public data more difficult.

The next time you request a public database, prepare to pay up.

That’s the message courts are sending to reporters and members of the public who regularly request copies of government databases as agencies are increasingly citing the proprietary nature of the software used in connection with such data to either make requesters pay large sums or cut off public access to the data entirely.

Two recent state supreme court decisions — and a similar case argued before California’s highest court in May — highlight how federal copyright laws and proprietary software used by state and local agencies and software developers are thwarting access to certain kinds of data.

Although the cases themselves dealt with geographic information systems data, commonly known as GIS data, transparency advocates worry that public agencies’ increasing use of private software and other proprietary systems to manage public data may close off access to broad categories of government records or require requesters to pay thousands of dollars to access the data.

By requiring requesters to pay high prices to access government data and sign licensing agreements that limit the use of it, agencies are effectively closing off access to public information, said Lisa Siegel, counsel for the Connecticut Freedom of Information Commission, who recently argued for access to copyrighted data in the state.

“It’s effectively access denied,” she said. “More than that, it puts control of access in the hands of the private party, rather than a public agency that is using those records.”

Conn. court cites copyright to withhold data

In Pictometry International Corporation v. Freedom of Information Commission, the Connecticut Supreme Court ruled in January that images and mapping software provided by private companies who claim copyright in the material put limits on whether governments can release such computer data under the state open records law.

The court ruled that mapping data maintained by the state’s Department of Environmental Protection was exempt from disclosure under the state’s public records law because a private entity had a copyright in the images used by the state.

The case began when Stephen Whitaker asked the agency for mapping and imaging data of the state, which the state obtained under an agreement with Pictometry. The agency refused to release the records, arguing that an exemption in the Connecticut Freedom of Information Act that allows agencies to withhold records pursuant to a federal law was applicable because the requested information was subject to the federal Copyright Act.

Although both the Connecticut Freedom of Information Commission, which enforces provisions of the state’s open records law, and a lower court agreed with Whitaker that the data was subject to disclosure, the Supreme Court ultimately upheld the agency’s refusal to release the records in a unanimous decision.

In upholding the agency’s refusal to release the records, the Supreme Court ruled that the federal Copyright Act preempted the disclosure requirements of the state public records law. In other words, the federal protections for the copyrighted material overruled the state’s command that the records must be disclosed.

The court also relied on a number of cases from other states that similarly ruled that copyright law could impose certain restrictions on copyrighted public records.

Siegel said that the case sets the state’s public records law back because it prevents public agencies from providing copies of records to requesters where it is subject to a private party’s copyright.

“This case could have a wide-ranging impact because it stands for the fact that the private third party can control the copying of these public records,” she said.

The case is also problematic because if the records are not subject to disclosure under the Freedom of Information Act, the only way a person interested in seeing the data can obtain it is to pay a license fee that many requesters simply cannot afford, Siegel said.

“They were going to charge him $25 per image where the rate under the (Freedom of Information Act) would have charged him 25 cents,” she said, referring to Whitaker.

The court’s decision affords businesses who provide materials to the government an additional form of protection, Siegel said. Many state records laws including Connecticut’s, already contain exceptions that allow businesses to withhold confidential business information they submit to agencies. Now businesses will be able to withhold non-confidential information if it is copyrighted, Siegel said.

Given how easy it is to copyright a work and the fact that private companies regularly provide a great deal of data to governments, Siegel said she worries that the decision will likely be applied to other government data as more agencies rely on private vendors to store and collect government information in databases.

“That’s really the new future in public records because there’s just so much data out there,” she said. “How do you control access to that? How do you charge for that? How do they redact?”

Copyrighted software is ‘inextricably intertwined’ with public records

The Ohio Supreme Court handed down a similar ruling this March in Gambill v. Opperman. The court ruled that an agency properly denied Robert Gambill access to Scioto County’s property mapping and imaging data because the county was not required under the law to separate the raw government data from the private, copyrighted software used by the county.

The court also ruled that although the county was not required to separate the records under the law, Gambill could obtain the data if he paid $2,000. That figure was the amount the county estimated it would cost it to extract the underlying data from the software.

“The engineer’s office cannot separate the requested raw data from the exempt Esri software files,” the court wrote in its per curiam opinion. “Therefore, consistent with other cases in which nonexempt materials are inextricably intertwined with exempt materials, the nonexempt records are not subject to disclosure under (the Ohio records law) insofar as they are inseparable.”

Justice Paul Pfeifer wrote an animated dissent in the case, arguing that requesters were being forced to pay for a government’s decision to use systems that make retrieving public data more difficult.

“The county engineer in this case has intertwined public records with proprietary software and expects citizens seeking public records to pay an exorbitant price to untie the knot,” Pfeifer wrote.

The ruling, Pfeifer wrote, harms public records requesters because they expect to pay reasonable fees for copying the records, not $2,000. He also expressed concern about the court sanctioning efforts by agencies to withhold public records by using private software systems.

“This case encourages public entities desiring secrecy to hide public records within a software lockbox and require individual citizens to provide the golden key to unlock it,” he wrote.

California set to decide whether public data is part of proprietary software

Attorneys argued a similar question before the California Supreme Court in May in Sierra Club v. Superior Court of the State of California, Orange County.

At issue in that case is whether GIS data is subject to disclosure under the California Public Records Act (CPRA). The conflict lies in a provision of the law that allows agencies to withhold computer mapping systems, including computer software developed by a public agency.

The Sierra Club is seeking access to county property map data, which provides details on the geographic boundaries of all land parcels in the county. The county offers to sell the database to the public for a fee, which can cost more than $10,000. Additionally, requesters have to agree to limits on how they use and distribute the information.

The county withheld the database under the CPRA, arguing that it could not separate the computer program from the database because the record was part of a computer mapping system that is exempt from disclosure.

Both a trial and appellate court ruled that the database could not be separated from the software and could therefore be properly withheld under the CPRA.

Rachel Matteo-Boehm, an attorney at Bryan Cave LLP who wrote a friend-of-the-court brief in the case on behalf of media parties, including the Reporters Committee for Freedom of the Press, said the case has big implications for requesters seeking access to data.

If the Supreme Court were to uphold the lower court’s ruling, it would mean that requesters seeking access to the data would have to pay thousands of dollars and also comply with restrictions on how they could use the data.

“It would basically say, ‘We have open records but only for those who can afford to pay,’” she said. “You just can’t effectively observe the government unless you can be on somewhat equal footing with the government on access to information.”

And allowing the government to attach strings to how requesters use the data in the form of licenses governing the access and use of particular databases would run counter to the CPRA, which does not allow agencies to inquire about the purpose of a request, Matteo-Boehm said.

Governments may be increasingly unwilling to provide data because they have invested a large amounts of time or energy into building and maintaining a computer system, Matteo-Boehm said. Also, agencies may sign agreements with vendors that put limitations on how agencies can disclose particular information, perhaps not realizing the potential conflict with public records laws’ disclosure requirements.

But Matteo-Boehm, who also litigated a similar case in another California appellate court, said using the complexities of the interplay between government data and the proprietary software systems used to analyze the data to withhold public records is a red herring.

“Software and data are distinct,” she said. “When you get into these sophisticated computer databases, I think it feels more complicated, but there is absolutely a distinction.”