The meeting took place in Oslo in the end of February. This is a summary of the presentations given. You can find the full presentations linked here.
The event was chaired by Annette Birkeland, director of The National Commission for the Investigation of Research Misconduct.
"Our mission today is to talk about the ethical challenges of open science", said Birkeland.
EU-defines Open science as 'a new approach to the scientific process based on cooperative work and new ways of diffusing knowledge by using digital technologies and new collaborative tools'.
Birkeland however underlined that openness has always been a core part of ethical norms in research and its guidelines.
"Openness is a principle in science. It's not a new notion, even if it's a buzzword right now", she said.
Ana Delgado, Associate Professor at the University of Oslo's Centre for Technology, Innovation and Culture, spoke on the moral economy of digital infrastructures. She said openness is an old principle in science, for instance experiments have taken place in the public since the 15th century. But our ways of producing and circulating knowledge has changed drastically, recently due to digital research infrastructures.
"This work of infrastructuring brings about a new moral economy of science which is openness" said Delgado. According to her, the digital era is one of the causes of the so-called reproducibility crisis in science.
"What is this crisis really about? I see it as an infrastructural crisis. A crisis about the means by which scientific knowledge is produced, shared and circulated."
But the digital research methods also offer solutions to the crisis, said Delgado. She used the field of biology as an example, where some forms of online publishing make it possible to click on a link connected to a certain DNA-sequence in a research paper and go directly to the dataset it is collected from.
"You have a way of codifying biological information in repository databases that you can then use in data models or design. This can bring about a means for collaboration, and it can give improvement of datasets back into the repository database. It can be seen as open source biology."
Delgado said this also changes the use of data.
"You can see communities emerging around data repositories. And there is a move from the encyclopedia to the Wikipedia way of production, a continuous process whose outputs are work in progress."
Open data challenges privacy
Grete Alhaug, senior advisor at the Norwegian Data Protection Authority (Datatilsynet) talked about the GDPR (the EU General Data Protection Regulation), which is about to enter into force in Norway.
"How does the principle of open data challenge the basic principles of privacy rights?" asked Alhaug. She said looking for answers in the Norwegian national strategy for data sharing, gave few answers.
"The strategy does not problematize how to solve this, it only states what we already know – yes, we must share data, but you must also consider privacy rights, so you just have to make sure."
Alhaug stated some basic principles on the processing of personal data:
• Fair processing, meaning you should be aware of what you are a part of, and collection and use of data should be understandable and adequate.
• That data should be collected for specified, explicit and legitimate purposes, and a consent for use must also be specific.
• Data minimization: the collection must be adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed. It must also be accurate and where necessary kept up to date.
Then there is the principle of anonymization, which Alhaug said is somehow challenged by the idea of open data. The GDPR-principle states that data should be kept in a form which permits identification of data subjects for no longer than necessary, given the purpose for which the data are processed.
"This means you should not save identifying data for longer than necessary, but what if you are sharing data? Must other researchers settle for anonymized data?"
A lot is still highly unclear regarding this new legislation and its relation to open data, concluded Alhaug.
"But the basic principles are still valid, and nobody says that will change".
Open access is here to stay
Ragnhild Ørstavik gave the next presentation, On openness in Publishing and Peer Review. Ørstavik is Associate Editor-In-Chief at the Journal of The Norwegian Medical Association. She is also a senior researcher at the Norwegian Institute of Public Health.
Ørstavik called open access 'a movement with a lot of feelings'. However, she said, it is not a matter of debating whether we shall have open access.
"That is like debating whether or not cell phones are a good thing. It is coming, and our government has told us this is how it is going to be."
Still it is important to remember that publishing costs money, said Ørstavik, and that means there is also money to be made in the industry. This has given rise to the so-called predatory journals, that we need to be cautious of.
"One of my friends sent an article to a journal named almost the same as the one she was supposed to send it to. It was published very quickly, and when she asked for a withdrawal, they charged her 20 000 NOK for it."
Open up the peer review process
We are all trying to adapt and establish new routines, explained Ørstavik. This also applies to peer review, in itself a rather new practice, following the former editorial review.
"Peer review is kind of the watermark of science. We need them", said Ørstavik.
A classic peer review is either double blind: author and reviewer are both anonymous, or single blind: reviewer is anonymous to the author.
"Ask scientists, and they will say that the double blinded peer review is the best for all parties, but it is often not possible in the Norwegian medical society. This leaves us with two options, single blinded and open."
Open peer review can mean several things, said Ørstavik.
• Author and reviewer are both identified
• Reviews are published
• All versions of the manuscript as well as reviews (and responses) are published
• Possibility of "post publication peer review"
Openness can make peer reviews more fair and transparent, it can allow scientists to take credit for their reviews and it can increase the learning potential, according to Ørstavik.
"It is a good idea. But there are reasons why we do not have them in the journal" she said, and stated some of the major concerns:
• Junior researchers may refrain from negative reports, because they fear it could affect their career negatively.
• Researchers may create «reciprocity strategies»: say something nice about my paper, and I will do the same for you.
• Reviewers might be even harder to find.
Open, but not naked
Stephan Lewandowsky continued the push to nuance the understanding of open science in his talk Being open but not naked: Balancing transparency with resilience in science. Lewandowsky is a psychology professor at the University of Bristol and the University of Western Australia.
"I am totally committed to open data. If we don't become transparent, others will do it for us" said Lewandowsky and gave examples of UK and US government actions on the issue.
"But we need a conversation about its full implications – because open data can be "weaponized". I want to show you that open data and transparency can be used for anti-scientific purposes" continued Lewandowsky. He pointed to a commentary in the journal "Big data and society" called "When open data is a Trojan Horse: The weaponization of transparency in science and governance", and then brought us to another example:
"Take the US data access Act from 1998 (and 2000). It states that all data from federally funded projects shall be made available. That "influential data" must be reproducible upon reanalysis by "qualified third parties". Sounds good, but then: "privately funded research is exempt from disclosure". The acts were drafted by the tobacco industry and its allies. If you give this industry any set of multi varied data, they can give you any fitting explanation for findings, other than tobacco being a health risk" said Lewandowsky.
This strategy, said the professor, can help industry or others create a favorable impression of scientific debate.
"The public calls for action when they think all doctors agree that tobacco is dangerous. They don't call for action when they believe there is scientific debate on a subject."
"Open data is political – it takes place in a political context. You can't be a scientist asking for transparency and ignoring that the tobacco industry, or any other industry, is asking for the same. And what motivates some to fight for transparency is not in the interest of transparency", said Lewandowsky.
Can't replicate a blowout
Another of his examples, is an act that would limit the US Environmental Protection Agency (EPA) to using only data that can be replicated and made publicly available for independent analysis. This might also sound fine? asked Lewandowsky. However:
"If an oil rig blows up, you can't replicate it. So, if all data must be replicable, it means they couldn't do research on such events."
"Do intentions matter, when considering open data? Up to a point no. But researchers lose control over their data when they are completely open – sometimes that matters" said Lewandowsky and referred to "The Belfast Project".
The project collected the oral history of 'the troubles' (the conflict in Northern Ireland from 1968-1998) – interviews with combatants from both sides. These are interviews with retired terrorists and killers. They were only to be released after they died, explained Lewandowsky.
"Boston college that held these tapes in a secure lock, handed them over to the Northern Ireland police after a subpoena, against the protest from the researchers. Sinn Fein leader Gerry Adams was arrested and held for questioning for four days in 2014. As a result, another oral history project had to end, because those who had consented would no longer share. If you promise confidentiality but can't uphold it, the consequences can be major."
Lewandowsky also offered some solutions to these issues. He believes in some kind of symmetry, as he put it.
"Why shouldn't people who ask for sensitive data follow the same rules and meet the same requirements as the person asking for the data to begin with? Everyone should have to preregister their intentions and conform to them. Why should we accept lowering our standards so that others can use our data as they please?"
He suggested open data should be considered in the peer review process by editors, not by authors who might be too involved. This could also prevent people from accusing scientist of holding back data, a strategy Lewandowsky said is being used to create an impression of dodgy science.
"A peer review should say 'this is the data, and this is exactly what it says'. On publication this could be confirmed by an independent body. If someone then accuses the scientist of withholding data by excluding some participants, for instance, the reviewer can say, 'yes, this is in accordance with the research protocol'."
Lewandowsky's final call to the audience was: "Don't let transparency damage science."