Final text of the update to the Public Sector Information Directive shows MEPs failed in moves to require publication of all publicly funded research. Rather than open by default, data will be “as open as possible, as closed as necessary”
The final text of the new EU directive aiming to put the results of all publicly-funded research into the public domain adds significant compromises to the European Commission’s original proposal, and rejects stronger measures to open up data that were demanded by the European Parliament.
The commission, parliament, and EU Council reached an agreement on an update to the current Public Sector Information (PSI) directive in January, and the council is expected to adopt it as law soon.
The final text has not yet been published, but a copy obtained by Science|Business shows that changes supported by the parliament, requiring all research funded by public money to be open by default, have been dismissed in favour of the Commission’s original proposal which makes allowances for criteria such as “confidentiality” and “legitimate commercial interests.”
Legislators have also added further caveats to address concerns about intellectual property, security and personal information.
In December last year, a consortium including the industry lobby group Digital Europe and the European University Association argued that measures to open up data from publicly-funded research risked chasing away investment by the private sector.
Research data is just one small but much fought-over part of the proposal, which updates existing rules intended to stop public sector organisations from putting restrictions on how others use their information once it is made public.
Data from publicly-funded research is dealt with in article 10: while the original version proposed by the Commission said member states must decide “open access policies” for such data, MEPs wanted the “open by default” rule, where the data would always be published unless there was an explicit reason why it should not be.
In the final version, the Commission’s wording has been preserved in its entirety, and a new sentence has been added saying, “Concerns relating to intellectual property rights, personal data protection and confidentiality, security and legitimate commercial interests, shall be taken into account in accordance with the principle ‘as open as possible, as closed as necessary’.”
The phrase “as open as possible, as closed as necessary” had always been there in the non-binding preamble to the legislation but lobbyists wanted to see it in article 10, and appear to have succeeded.
Though there is very little opposition to the phrase as a general principle, making it legally binding begs the question of what exactly will be necessary under the law.
The final version also says that the data must be published according to what are known as the FAIR principles: Findable, Accessible, Interoperable (that is, can be used across different systems and software), and Reusable.
Commission sources said they are happy with the outcome.
Julien Chasserieau, policy officer at Digital Europe, a trade association representing tech companies including Google and SAP, said the final version, “strikes a great balance and addresses the issues we highlighted in our joint statement in November.”
Potential clash with copyright regulations
However, not all of the co-signatories to last November’s joint statement are completely satisfied. Lidia Borrell-Damián, research and innovation director at the European University Association, said while she agrees with the basic principle that publicly-funded research data should be open, the changes to the PSI Directive might clash with another fought-over piece of legislation on copyright.
“We actually believe that this directive should not deal with research data, because it’s the Copyright Directive that has an article on research data and text and data mining,” said Borrell-Damián. “The public sector information [directive] should actually, for the sake of coherence, refer to the copyright regulation, because we don’t really understand why this has to be addressed again.”
The Digital Single Market (DSM) Copyright Directive is not law yet, but there has been intense lobbying over the legality of using search tools to do text and data mining of a large number of journals at once.
The legality of algorithm-driven text and data mining is in doubt because it can involve copying protected information. As it currently stands, the DSM Copyright Directive would create an exemption for academic researchers only, allowing them to use text and data mining tools on lawfully-accessed content for educational, non-commercial purposes.
Borrell-Damián suggests the new PSI Directive could render this limitation irrelevant by opening up data to anyone. “There is a risk [that] while in the other copyright regulation things are very well protected and established, [and] the interests of research and education are very well protected, here we have this openness.” She stressed that she was highlighting a risk, not making a prediction. What happens will come down to how article 10 gets interpreted and implemented.
Once adopted, the update will change the name of the PSI Directive to the “Directive on Open Data and Public Sector Information.” Though this change was not in the commission’s proposal, parliament added it at the commission’s request.