The talk will offer a comparative review of policy-making in the area of research impact evaluation in UK (REF), Poland (EJDD) and Norway (Humeval and Sameval). Poland and Norway have used an ex-post, expert-review system modeled on the British REF. There are several analogies between the studied impact evaluation systems, including similar definitions of impact, the use of case studies as the basis for evaluation, the structuring of the impact template and English as the language of evaluation. There are also several differences: the mode of introduction of the exercise (gradual vs. shift), whether the exercise is tied to funding, and the level of transparency of the policy-making and evaluation process. The text will present an overview of the three approaches on impact evaluation and attempt to answer the following question: 1) How does the articulation of research impact change depending on the goals of the exercise and the broader academic and social context into which the exercise is introduced; 2) Consequently, how do the effects of the exercise differ from one national context to another?

After longer comparative part of the talk (based on an article submitted for review to Research Evaluation), I will attempt an interpretation of the ongoing process of policy-borrowing in the area of impact evaluation in terms of centre-periphery dynamics (as described by Immanuel Wallerstein) and/or Antonio Gramsci's theory of hegemony. The goal of the talk will be to test out theoretical theoretical perspectives on the data collected. I welcome any remarks and recommendations from the audience.

The talk explores the epistemic dynamics catalysed by researchers advocating for the clinical relevance of environmental epigenetics in psychiatry. I do so based on an in-depth literature analysis of peer-reviewed research articles and interviews with researchers who conduct epigenetic research in psychiatry. In demonstrating how relevance builds a crucial yet ambivalent bridge between basic research and clinical application, I explore tensions arising in relation to the acceptable level of uncertainty for epigenetic knowledge to be considered relevant. I further trace how epigeneticists aim to counteract emerging problems to their claims about the clinical relevance of epigenetics through performing interdisciplinary, big-data research. Finally, I show that, nonetheless, certain epistemic problems persist and discuss both their roots in the specific epistemic history of psychiatric epigenetics as well as in the systemic pressures to promote relevance early on in emergent research fields. With this talk, I contribute to STS scholarship that explores how modes of relevance feature in different scientific domains. At the same time, my talk contributes to a better understanding of how environmental epigenetics is adopted and adapted in different research fields within biomedicine and how field-specific norms, infrastructures, and societal expectations affect its uptake, articulation, and epistemic development.

Recent scholarship has described the “reproducibility crisis” and its associated reform movement as a social movement or a scientific-intellectual movement. This talk will argue for an alternative framing of these events which de-emphasizes high-status intellectual actors and their agendas for change, and emphasizes instead structural aspects and how they shape which reforms come to be seen as possible and desirable.

Taking this alternative lens, we could see the reproducibility movement not as a successful network of methodologists stitched together by Twitter and investment from private foundations, but as one instantiation of a broader diffusion of a “regulatory ethos” from highly regulated contexts into academic settings. Together with Lara Keuck, we describe the regulatory ethos as a specific type of scientific ethos (following Merton) that relies on documentation practices and favors plans over situated actions, uniformity over heterogeneity, and validation over external validity.

Through a detailed history of the emergence of the reproducibility crisis in American biomedicine, I will show how pharmaceutical drug development and clinical medicine – two highly regulated spheres of practice – were key to providing the evidence and solutions that first generated support from key figures at the National Institutes of Health for reproducibility reform. From this I argue that this reform movement should be understood analytically as not merely an expression of the open science ideals of its most charismatic leaders, but as part of ongoing efforts to ensure the interoperability of experimental systems across distributed research spaces.

Research guidelines and the scientific literature in general are full of ideas and recommendations of how proper science should look like. However, it remains an open question how the actual reality of research in the sciences relates to notions of proper or responsible science in, for instance, European research integrity guidelines? To answer this question, I conducted an ethnography of cognitive sciences in five cognitive science labs in Germany, the Netherlands and the UK. The aim is to understand how and under what conditions knowledge is produced and whether ideas of proper conduct that can be found in guidelines and the literature actually have a place in research reality. Hence, it is about capturing the perspectives of the researchers who are supposed to live in compliance with such guidelines and recommendations. To this end I further inquired what researchers think about how and whether current recommendations for the improvement of science fit into their everyday lives as researchers by conducting semi-structured interviews. Put shortly, I observed a complex mixture of alignment and mismatch between notions of good science and the research realities.

In the case of artificial intelligence, hyperbolic predictions of the emergence of intelligent machines, even ‘super intelligences’, consist of both dystopian fears of human suppression and extinction, and utopian hopes of human flourishing through freedom from labor and illness as well as unparalleled economic growth and prosperity. At the heart of the controversies between these two, we argue, are emergent and conflicting assumptions about what it means to be human, or rather, what defines humanness.

To address this topic, of how the understanding of humaness is constructed in relation to AI and how the (future) agency of AI and Humans are imagined,  we turn to the genre of popular science and the imaginaries of the possibilities and effects of a future in which intelligent machines have bypassed many human capacities.  Popular science as a genre is interesting in its ambition to translate inter-academic knowledge production about AI development while at the same time dramatizing it and making it relevant for business, politics, and the public.

First, the chapter deconstructs the imaginaries of a future shaped by super intelligent AIs and discusses how this imagined future builds on particular and narrow definitions of humanness - as essentially biological cognitive processors, but also as distinguishable as creative/non-creative and neuro-typical and neuro-diverse/passive and active. Secondly, we turn to the construction of AI as a “floating signifier” an object, a thing, that is devoid of meaning.

This paper is accepted as a chapter to the upcoming Routledge book: The inner World of Artificial Intelligence, and the section “Hidden and Subsumed Humans in Artificial Intelligence.