The Ethics of not releasing Data #1: the context

Source NYMag.

I have been funded to answer biological questions around the safety of nanomaterials. In particular, to what extend can we understand and therefore predict the safety of new materials, often referred to as safe-by-design. We started with eNanoMapper and continued with NanoCommons, NanoSolveIT, RiskGONE, and SbD4Nano, and supervising a PhD student working on the EU-ToxRisk and OpenRiskNet projects. Here, "we" includes in chronological order Cristian, Bart, Freddie, Linda, Nuno, Serena, Laurent, Luc, Jeaphianne, and Ammar. And of course with all the other partners in the fascinating projects. Oh, and via the new ELIXIR Toxicology Community with the ELIXIR projects too.

For integrative systems biology research at BiGCaT we need knowledge to analyze data. And there lies the culprit: only a small percentage of knowledge does not require reading a PDF, manually transcribing the knowledge. Ergo, the eNanoMapper project. One recurrent theme is the lack of useful data. There are many angles here, and this blog series starts collecting information about just only one angle. The ethics of not releasing data.

However, I am not a full-time philosopher. It's just that my hands are dirty of working with and around this moral or ethical question. There we have a first aspect of this one angle: is the decision to release data or not a moral or ethical decision?

Over the past 27 years the topic first came up just as an inefficiency. There was research I could easily do because data was either proprietary (and I was not rich enough), did not give me enough detail (low quality, commercial databases), and often did not allow me to talk about my research (which is resharing modified data). But hey, I was just a student and capitalism is ethically right decision.

In fact, there are some arguments in favor of this world view. The arguments that get repeated again and again are that academic research setting is not suited to disseminate knowledge. Publishers, companies, etc, should sell data and use that money to maintain the code. Basically, it says that universities libraries are not suitable to share scientific knowledge. Well, that last line sheds some light into what I think of this argument. In fact, this is the whole nature of the discussion about the sharing or not sharing the vaccine IP. One caveat, patented data actually is open and you can modify and reshare it. Patent law is just like the Non-Commercial clause in Creative Commons licenses: you cannot make a business out of it, without getting a different license from the patent holder.

Now, over the years I saw more and more arguments pass by. Like the declaration of human rights, that explain everyone has the right to benefit from new scientific knowledge. Or the sad stories that people die because the information was not in the right place at the right time.

This latter is the context of the exploration I start in this series. Again, I am not a full-time philosopher and I will be looking for key literature, arguments, etc, to understand the theoretical foundation and (in)validity of this statement:

It is unethical to not make data available for research.

Some starting thoughts. We waste a lot of rare research money:

• to repeat experiments for data that already exists
• to make existing data useful in research (this is what FAIR is about)
• because data is not disseminated enough, people die
• because data is not readily available, it limits new innovation
• governance needs to be more transparent, and this requires access to underlying data, as that is the ultimate arguments supporting the governance

Second, arguments against the claim include:

• quality assurance of open data is not possible (yeah, people really use this argument)
• maintenance of data in libraries is too costly to be done by universities
• hey, capitalism is good, not bad

I guess it is obvious why this exploration is warranted, after reading my short list of points, right?

I prefer replies as blog posts, journal commentaries, etc. Please do cite this blog post when you reply.