Science would not be possible without trust among experts, trust of the public in experts, and reliance on scientific instruments and methods. The rapid adoption of scientific foundation models and their use in AI agents is changing scientific practices and thereby impacting this epistemic fabric which hinges on trust and reliance. Foundation models are machine learning models that are trained on large bodies of data and can be applied to a multitude of tasks. Their application in science raises the question of whether scientific foundation models can be relied upon as a research tool and to what extent, or even be trusted as if they were research partners.

Conceptual clarification of the notions of trust and reliance in science is pivotal in the face of foundation models. Trust and reliance form the glue for the increasingly distributed epistemic labour within contemporary technoscientific systems. We build on two concepts of trust in science, namely trust in science as shared values, and trust in science based on commitments to processes that provide objective claims. We analyse whether scientific foundation models are research tools to which the concept of reliance applies, or research partners that can be trustworthy or not. We consider these foundation models within their socio-technical contexts.

Allocation of trust should be reserved for human agents and the organizations they operate in. Reliance applies to foundation models and artificial intelligence agents. This distinction is important to unambiguously allocate responsibility, which is crucial in maintaining the fabric of trust that underpins science.@en

Innovations in biotechnology increasingly shape our societies and our planet. The stream of innovations that could be witnessed in the past few decades opens up new ways to do agriculture, to provide healthcare and to produce compounds and materials, amongst many other things. Many of these innovations rely on biological data. The ability to extract a plethora of biodata vastly increased over the past few decades. These biodata provide deeper insights into the workings of biological systems, thus constituting a fertile ground for bio-inspired innovations. For example, population genomics data provides the basis for a personalized health care. Biodiversity data derived from ecosystems provides the basis for the identification of novel drugs, high value chemicals and materials. Biodata is increasingly crucial when aiming for a flourishing bio-economy and biomedicine. Biodata-based innovations thereby raise very substantial ethical questions. Pronounced cases like human genome editing, or the engineering of entire species via gene drive technologies, make clear that innovations need to go hand in hand with societal deliberation and an ethical accompaniment of technology development. The question though is how such responsible innovation can be organized. Extraction of biodata is done at a speed that surpasses Moore’s law. And the resulting biodata-based innovations are fast-paced. It is therefore highly needed to consider how a responsible guidance of innovation in biotechnology can be accomplished, in view of the biodata-avalanche. This question provides the entry point for this dissertation. Central to this analysis is the special ontological and epistemological position of biodata. Biodata resides at the interface between the biophysical world and the realm of human language and meaning. This makes biodata a central locus when pursuing a value-driven accompaniment of innovation in the field of biotechnology. @en

Synthetic biology, as an engineering approach to biological systems, has the potential to disruptively innovate the development of vaccines, therapeutics, and diagnostics. Data accessibility and differences in data-usage capabilities are important factors in shaping this innovation landscape. In this paper, the data that underpin synthetic biology responses to the COVID-19 pandemic are analyzed as positional information goods—goods whose value depends on exclusivity. The positionality of biological data impacts the ability to guide innovations toward societally preferred goals. From both an ethical and economic point of view, positionality can lead to suboptimal as well as beneficial situations. When aiming for responsible innovation (i.e. embedding societal deliberation in the innovation process), it is important to consider hurdles and facilitators in data access and use. Central governance and knowledge commons provide routes to mitigate the negative effects of data positionality.

@en

Digitalization of biological populations and ecosystems changes our relation towards them. In silico representations of natural systems make them available as resources that allow for novel ways of deriving economic value. These extracted data and models also open novel routes for responsible innovation based on biological systems and derived biological data. Responsible innovation based on natural resources is explored using the common pool resource framework and using the emerging field of biodiversity sequencing as an example. Natural syste­ms that have a vast digital representation which is shared by a community have aspects from both a natural resource commons and from a knowledge commons, but differ in their structure and dynamics. We therefore propose the concept of ‘Twin Commons’: the institutional arrangement of natural resources that have a tightly linked digital component which is shared and governed by a community, and that have research and innovation as important outlets.

@en

Personalized medicine uses fine grained information on individual persons, to pinpoint deviations from the normal. ‘Digital Twins’ in engineering provide a conceptual framework to analyze these emerging data-driven health care practices, as well as their conceptual and ethical implications for therapy, preventative care and human enhancement. Digital Twins stand for a specific engineering paradigm, where individual physical artifacts are paired with digital models that dynamically reflects the status of those artifacts. When applied to persons, Digital Twins are an emerging technology that builds on in silico representations of an individual that dynamically reflect molecular status, physiological status and life style over time. We use Digital Twins as the hypothesis that one would be in the possession of very detailed bio-physical and lifestyle information of a person over time. This perspective redefines the concept of ‘normality’ or ‘health,’ as a set of patterns that are regular for a particular individual, against the backdrop of patterns observed in the population. This perspective also will impact what is considered therapy and what is enhancement, as can be illustrated with the cases of the ‘asymptomatic ill’ and life extension via anti-aging medicine. These changes are the consequence of how meaning is derived, in case measurement data is available. Moral distinctions namely may be based on patterns found in these data and the meanings that are grafted on these patterns. Ethical and societal implications of Digital Twins are explored. Digital Twins imply a data-driven approach to health care. This approach has the potential to deliver significant societal benefits, and can function as a social equalizer, by allowing for effective equalizing enhancement interventions. It can as well though be a driver for inequality, given the fact that a Digital Twin might not be an accessible technology for everyone, and given the fact that patterns identified across a population of Digital Twins can lead to segmentation and discrimination. This duality calls for governance as this emerging technology matures, including measures that ensure transparency of data usage and derived benefits, and data privacy.@en