Seeing the bigger picture: who benefits from data?
Our first The Bigger Picture webinar on 25 June focused on the governance and ethics of digital data and AI, with Professor Barbara Prainsack from the University of Vienna.
Whether its researchers using remote sensing data or scientists using AI to create decision-making tools - how can we ensure that new technologies enhance rather than undermine our capacity to live well on the planet? And why is it that there are hundreds of AI ethics guidelines, but studies show that these guidelines have had very little impact on actual practices so far?
Professor Prainsack explained that one of the reasons for this lies in the categories that are used to govern the ‘molecules’ of AI, namely digital data. Digital data governance is stuck in the paper age, focusing mostly on individual-level control over personal data and the idea that benefits and risks are mostly tied to data types. Data governance approaches that are appropriate to the digital era don’t focus on data types, but instead on the purposes for which data are used. Who benefits from it, who bears the cost? Moreover, data governance should acknowledge the relationality of data – namely that they come out of, and represent, relations between people, or between people and non-humans.
Who controls data, who is visible and who is invisible?
One of the main reasons people don’t trust data users is that they think that the benefits are not shared with the people that the data comes from. People are aware that their data are used, but feel they have little say in how this takes place, leading to a disconnect, a legitimacy gap. Environmental data, for example, can trigger policy interventions and consequences for communities, even though people didn’t specifically consent to being part of a dataset.
Professor Prainsack called for a rethink of the governance of data, to consider data as a jointly owned resource. Data solidarity is a new approach to ethical data governance, which acknowledges that data is relational, and can impact entire communities and ecosystems. It is focused on three main pillars: supporting high-public-value-uses (eg climate, health, education); preventing and mitigating harm (eg bias, unfair profiling); ensuring fair returns (benefit sharing).
A public value assessment tool (PLUTO) has been developed, to try and implement this in practice. (See https://pluto.univie.ac.at/)
Stay tuned for our next The Bigger Picture webinar on 20 October with Professor Nikolaus Forgó from the University of Vienna, which will focus on artificial intelligence, intellectual property and data protection.
