Artificial Intelligence, Human Agency,
Political Action

Artificial Intelligence, Human Agency, Political Action

September 30th 2024

Programme

10:00 Opening words (Thomas Joseph White OP, Rector of the PUST / Helen Alford OP, Dean of the Faculty of Social Sciences, PUST)

10:15 Introduction (Gábor L. Ambrus, lead researcher of the project ‘Human Freedom in the Age of AI’)

10:45 ‘AI and Democracy: Political and Epistemic Agency’ (Mark Coeckelbergh, Univ. Vienna); response (Helen Alford OP); discussion

11:45 ‘Resistance Is Futile? Technological Change & Human Agency’ (Henrik S. Saetra, Univ. Oslo); response (Frantisek Stech); discussion

12:45 Lunch

14:15 ‘Moral Dilemmas Surrounding AI-Enabled Systems in Society and on the Battlefield’ (Gregory M. Reichberg, Peace Research Institute, Oslo); response (Giuseppe Casale); discussion

15:15 ‘AI For Political Organisation: the Case of Intra-Party Democracy’ (Claudio Novelli, Un. Bologna); response (Margherita Daverio); discussion

16:15 Break

16:45 Roundtable discussion with all speakers / conclusions

17:45 End

The Faculty of Social Sciences at the Pontifical University of St. Thomas Aquinas (Rome) and the Centre for Digital Culture of the Dicastery for Culture and Education (Vatican), in cooperation with the Theology and Contemporary Culture Research Group at the Charles University (Prague), cordially invites you to the research seminar entitled
‘Artificial Intelligence, Human Agency, Political Action’.

With 2024 gathering momentum – a year which has been labelled ‘the biggest electoral year in history’ –, there is a considerable apprehension in public discourse about the possible harmful impact of technology on politics which may exceed what happened in 2016 at the U.S. elections and the fateful Brexit referendum in the U.K. That year, undemocratic campaign methods, although having been in the making for several years, reached an unprecedented level: with the help of vast data collection and user profiling enabled by social media, voters were targeted and influenced at scale with messages designed to manipulate them towards a specific political outcome. Meanwhile, recent advances in artificial intelligence have further exacerbated the situation in which what is at stake is nothing less than the very foundations of liberal democracy:  the deepfakes and the plethora of misinformation potentially manufactured by AI can circumvent vital ingredients of the democratic process like well-informed deliberation and conscious decision-making.

This looming possibility among the many ways AI can be implemented in society clearly indicates that the well-established discourse on the ‘ethics of AI’ needs to be combined with a broad discussion about the ‘politics of AI’ which could draw on both political philosophy and the philosophy of technology (Coeckelbergh 2022). In the philosophy of technology – partly due to the ongoing influence of Marxist thought –, it has long been recognised that technological innovation can be seen as a force in society with enormous political significance. Technology itself is in the position to make history (Heilbroner 1967); technological artefacts themselves can necessitate layers of social organisation around them – sometimes irrespective of the intentions of their inventors and designers (Winner 1986). Amid such plausible insights, however, there remains one decisive question: while it makes sense to assume that technology ‘acts upon’ politics and society, is there a possibility – and, if so, under what conditions and to what extent – of politics ‘acting upon’ modern technology in general and artificial intelligence in particular? 

As a matter of fact, this question is to be raised against the background of another one concerning the truth or untruth of a philosophical position on technology: determinism. Techno-determinism does not only mean technology running its own autonomous course, but also its determination of everything else in society; indeed, its autonomy amounts to the heteronomy of everything else subjected to it (J. Ellul, L. Winner). A particular kind of technological determinism arose in two phases in the 1960s and the 1980s, premised upon the deterministic impact of technical media defining human perception, mentality and consciousness (M. McLuhan) and, what is more, having human consciousness as its mere effect (F. Kittler). When it comes to contemporary times, determinism is claimed to be widespread among technologists and academics in STS (science and technology studies) (Wyatt 2007), coupled with a particular creed in the IT sector called ‘inevitabilism’ which professes the necessity of each phase of technological innovation, no matter who carries it out (Zuboff 2019).

In a deterministic scenario, human agency, including political action, is by definition devoid of any freedom. However popular techno-determinism may be among technologists and academics, this view cannot ever be a viable option among politicians, decision makers, regulators or activists. Yet, to embrace the possibility of free and democratic political action on technology is not to deny the vulnerability of this free action when faced with the challenge of artificial intelligence. AI has a certain ‘shadowy character’ unparalleled by technologies from the past. The basics of nuclear energy and nuclear power plants, for instance, have a ‘visibility’ and can be explained with relative ease, and their implications can be made clear without much difficulty to the average citizen. This is hardly the case with AI. Knowledge about AI and expertise in AI have a somewhat arcane quality, barely accessible to the uninitiated. Clearly, knowledge plays a pivotal role here: if the political agency of citizens and their democratic participation in the debate on AI need to be free, this freedom requires them being well-informed. Knowledge increases, while the lack of it decreases freedom.

To name another factor with influence on the freedom of human agency and political action concerning AI, there is a possible ‘window for free action’ in the implementation of every momentous technological innovation which can be called a ‘historical juncture’: at the beginning when the technology has as yet not been broadly implemented, it can be more easily acted upon than at a later stage when it gathers something like a growing ‘gravitational pull’ or ‘path dependence’ with political actors’ free manoeuvring being increasingly restrained (cf. Coeckelbergh and Saetra 2023). Undeniably, when it comes to AI, we are at such a historical juncture, and political debate and action on AI is a matter of urgency.