Jonathan Jeschke, Sophie Lokatis, Isabelle Bartram, Klement Tockner

We propose the concept of knowledge in the dark – or short: dark knowledge – and outline how it can help clarify why in our current era of Big Data, the knowledge (i.e. evidence-based understanding) of people does not seem to be substantially increasing despite a rapid increase in produced data and information. Key reasons underlying dark knowledge are: (1) the production of biased, erroneous or fabricated data and information and (2) the inaccessibility of data and information, both for example due to sociopolitical or financial interests; (3) the incomprehensibility of data and information, for instance due to specialized jargon and complex disciplinary knowledge; and (4) the loss of previous knowledge due to, for example, the extinction of scientific disciplines. OECD data show that only a fraction of the global investments into research and development are done with the principal intention to increase public knowledge. Most investments are by the industry, steadily increasing with time, whereas governmental investments have been decreasing in relative GDP terms. This pattern suggests an increasing privatization of knowledge. But also in the academic realm, where increasing public knowledge is a primary aim, several factors lead to dark knowledge. We highlight four of these factors – loss of academic freedom, research biases, lack of reproducibility and the Scientific tower of Babel – and offer ways forward to tackle them, for example establishing an international court of arbitration for research and developing advanced tools for research synthesis.