Past Talks

Does consciousness require biology or can systems made out of other materials be conscious? I develop an argument for the view that it is (nomologically) possible that some non-biological creatures are conscious, including conventional, silicon-based AI systems. It assumes the iterative natural kind (INK) strategy, according to which one should investigate consciousness by treating it as a natural kind which iteratively explains observable patterns and correlations between potentially consciousness-relevant features. The argument is based on the insight that we can already anticipate that future developments would give us reasons to attribute consciousness to some non-biological creatures. According to the argument, an idealized scientific investigation – based on the INK strategy – would deliver the result that some possible non-biological creatures are conscious, and the outcome of such an ideal application corresponds to what is actually the case. My argument for the former premise is based on the claim that theoretical virtues and pre-theoretical principles support attributing consciousness to psychological duplicates, i.e., non-biological, silicon-based creatures which share the coarse-grained functional organization of humans.

The question whether a machine – a computer, a robot or any other form of artificial system – could be sentient is certainly entertaining, no end of science fiction deals with the question and sometimes very engagingly. But why is the question of artificial sentience (or “awareness”, or “consciousness”) raised in science and why invest public funding in this research? Is conscious AI at all possible, or even desirable?

Here is a very popular view on what user rational trust in AI requires: The Explanation View of AI Trust: User rational trust in AI requires an explanation of why the AI has reached the conclusion it has. We think that The Explanation View of AI Trust is wrong. It’s not true of trust in general that rational trust (even typically) requires understanding why, and it’s not the case that AI communication generates any special normative requirement that there should be an explanation why that grounds rational trust. This doesn’t mean that we think there is nothing to be gained by XAI—we prefer explainability all else being equal! But understanding how to increase trust (when appropriate) in AI requires the right diagnosis. In order to understand how to increase trust in AI, we think it’s better to focus not on AI explainability but instead on AI trustworthiness. That is, in this talk I will defend what we call the Simple View: The Simple View of AI Trust: User rational trust in AI requires AI trustworthiness.

From the perspective of artificial intelligence application, law is at a crossroad—on the one hand, it is an ideal application because it is entirely based on written text and therefore suitable for methods based on LLMs. On the other hand, however, law is bound by a number of legal, ethical, and other rules for its application, and the use of artificial intelligence sheds new light on these rules. I will present this paradox of the legal domain using the example of the Constitutional Court case law search engine that I am developing. The key question here is: how should assistance tools for legal practice work in order to be effective and lead to more accessible law, but at the same time not pose a risk to the protection of fundamental rights?

Consuming information online today means dealing with AI-powered technologies from filter algorithms to deepfakes on a daily basis. The bells of the so-called infopocalypse or epistemic apocalypse have been sounded and are being discussed both inside and outside academic settings with claims that the inherent opacity of these technologies could harm our epistemic practices. This paper challenges two narratives: that epistemic opacity is primarily a technological problem requiring technical solutions and that it represents a fundamental threat to human epistemic agency that must be overcome. In particular I argue that the necessity of using AI-mediated technology to stay informed forces us to adopt a r(el)ational approach to ignorance: we must strategically choose what not to know (rational ignorance) while negotiating how we depend on others through technologically mediated epistemically opaque channels (relational ignorance).

Does linguistic reference to the external world require a basis, or grounding in, perception? Do today’s large language model (LLM)-based chatbots have such perceptual grounding? In this talk, I give a qualified “yes” answer to the first question, focusing on a notion of reference as a direct relation to reality and a broad notion of perception. I do not answer the second question definitively, but I propose that the answer depends on whether LLM chatbots’ processing of linguistic input itself can provide the needed perceptual basis. I make a start on assessing this by considering whether and how prior linguistic uptake serves to ground human referring, especially in the broadly Kripkean chains of communication that have also been appealed to recently to explain reference by LLM chatbots.

This paper engages with fictionalist accounts of verbal human-AI interaction, according to which, although lay AI-users know that AIs are not intentional agents (and hence that their linguistic output is strictly speaking meaningless), laypeople pretend/imagine that they are intentional agents when talking to them. Reviewing the relevant experimental philosophy literature, I suggest that a substantial group of lay AI-users — the uncaring users — engage in bullshit action: They do not know, and, more importantly, do not care whether AIs are intentional agents. Still, they act as if this is the case when talking to them.

The uncanny entities in the AI “suitcase” increasingly straddle the border between person and thing. Western legal systems however have for two millennia relied on a strict, tertium non datur person-thing dichotomy. Their resistance to change recently took the form of the AI Act, in which AI systems are treated as things, products on the internal market, despite earlier calls from the EU Parliament for the attribution of legal personhood to the most complex ones. I argue that there are multiple ways in which AI systems escape the binary categorization as either things, due to their anthropomorphic features, or persons, due to the imperfect overlap between their capacities with human ones. This mismatch, if ignored in jurisprudence, creates gaps in liability attribution and legal protection generally.

I propose that the capacities setting AI systems apart could amount to legal agency. The concept of legal agency employed here is different from the law of agency. It does not only refer to contractual legal relations of representation, but to a legal status found somewhere in-between legal personhood and thinghood, or rather beyond these. Legal agents do not quite overlap with legal persons, lacking the complete, robust array of rights and obligations that persons hold, but they would not fall into the passive category of mere things that legal relations are about either. At the same time, not all legal persons are legal agents, such as infants, who have rights but lack the capacity to act. Legal agents have the capacity to act with legal effect, that is, to perform acts and enter legal relations (e.g., entering contracts, acquiring rights, incurring liabilities) that the legal system recognizes as legally valid and binding.

However, for AI systems to be jurisprudentially recognized as legal agents, they need to be considered the source of an action. The standard theory of agency in philosophy associates this with mental states, as does the legal theory and philosophy of law. Thus, from this realist position, it is argued in philosophy that an AI system may not have internal states to ground the ascription of representational mental states, thus being incapable of intentional agency. Assuming that law follows closely this philosophical account, there are two avenues for artificial legal agents then: either display (something akin to) mental states or then comply with a set of less demanding requirements that “naturalize” agency. As for the first, the sparse accounts of AI agency in legal literature have so far relied either on the “sense-think-act paradigm” coming from the technical domain to describe the internal states of these artificial agents that mimic mental states, or on  what is known in the philosophy of action as “the intentional stance”, which involves treating AI agents as if they are intentional entities as a strategy to deal with their actions and without looking into their make-up.

As to the second, the law lags behind philosophy, without concepts such as agency without mental representations in its toolbox. I propose changing that and replacing “Liberal” agency, which has been the only legal concept thereof for the better part of two centuries. Since persons with disabilities, animals, and now AI systems seem to mark the end of the individual, highly rational, autonomous person as the only legal agent that can be taken into account, this contribution will thus flesh out a concept of legal agency that is mindful at least of the seeming mindlessness of AI.

How do we ensure that AI systems remain aligned with human values? Recently, worries have been expressed that the autonomy and increased ‘intelligence’ of AI systems may lead to systems that get out of control and are harmful to humans. This has led to proposals for AI systems that ensure value alignment by tracking human values or preferences in real-time (for example, Stuart Russell in his book Human Compatible: Artificial Intelligence and the Problem of Control). In my talk, I will criticize such proposals for being insufficient for addressing the alignment problem.

Generative AI models are now being used to create synthetic climate data to improve the accuracy of climate models, and to construct virtual molecules which can then be synthesized for medical applications. But generative AI models are also notorious for their disposition to “hallucinate.” A recent Nature editorial defines hallucination as a process in which a generative model “makes up incorrect answers” (Jones, 2024).  This raises an obvious puzzle. If generative models are prone to fabricating incorrect answers, how can they be used responsibly? In this talk I provide an analysis of the phenomenon of hallucination, and give special attention to diffusion models trained on scientific data (rather than transformers trained on natural language.) The goal of the paper is to work out how generative AI can be made compatible with reliabilist epistemology. I draw a distinction between parameter-space and feature-space deviations from the training data, and argue that  hallucination is a subset of the latter. This allows us to recognize a class of cases in which the threat of hallucination simply does not arise. Among the remaining cases, I draw an additional distinction between deviations that are discoverable by algorithmic means, and those that are not. I then argue that if a deviation is discoverable by algorithmic means, reliability is not threatened, and that if the deviation is not so discoverable, then the generative model that produced it will be relevantly similar to other discovery procedures, and can therefore be accommodated within the reliabilist framework. 

In this talk I address the extent to which digital and robotic generative AI avatars that represent individual persons complicate responsibility gaps opened by increasingly autonomous AI systems. I argue that using GenAI avatars requires us to lose some level of agency in terms of control and knowledge, which are precisely the two main criteria widely used to ascribe moral responsibility. Use of digital and physical GenAI avatars therefore opens new responsibility gaps, which refer to the exact nature of the relationship between the human users and their avatars powered by generative AI, and which can adequately be called “proximity gaps”.

 Recent work in philosophy of technology has come to bear on the question of responsibility gaps. Some authors argue that the increase in the autonomous capabilities of decision-making systems makes it impossible to properly attribute responsibility for AI-based outcomes. In this article we argue that one important, and often neglected, feature of recent debates on responsibility gaps is how this debate maps on to old debates in responsibility theory. More specifically, we suggest that one of the key questions that is still at issue is the significance of the reactive attitudes, and how these ought to feature in our theorizing about responsibility. We will therefore provide a new descriptive categorization of different perspectives with respect to responsibility gaps. Such reflection can provide analytical clarity about what is at stake between the various interlocutors in this debate. The main upshot of our account is the articulation of a way to frame this ‘new’ debate by drawing on the rich intellectual history of ‘old’ concepts. By regarding the question of responsibility gaps as being concerned with questions of metaphysical priority, we see that the problem of these gaps lies not in any advanced technology, but rather in how we think about responsibility.

While LLMs’ capacity for semantic understanding has been widely debated, less attention has been paid to whether they or other AI systems can perform speech acts. The speech act of assertion involves not only producing outputs with descriptive functions, but also making substantive commitments to the aptness (perhaps truth) of these outputs, according to the norms of assertion. This entails that only entities that can be sanctioned for breaching these norms can make assertions. In ‘AI Assertion’ (with Emanuel Viebahn), we argued that this means that current AI systems cannot assert. In this talk, I will present our arguments and briefly consider whether anything has changed since we wrote the paper.
Link to ‘AI Assertion’: https://osf.io/preprints/osf/pfjzu

With the digitalization of society, the debates and research efforts relating computational systems with regulations have been widely increasing. Yet, most arguments and solutions refer to established computational/formal frameworks, rather than targeting more fundamental mechanisms. Aiming to go beyond this conceptual limitation, I will elaborate on taking “normware” as an explicit additional stance — complementary to software and hardware — for the interpretation and the design of artificial devices, highlighting the opportunities of a normware-centred engineering, as well as the problems it brings to the foreground.

Flipping through the major journals in the ethics of technology, one gets the impression that the rise of AI is an ethical catastrophe. The big debates in AI ethics almost invariably revolve around problems, while the positive aspects of AI are rarely talked about. Among ethicists, there is a ‘rising tide of panic about robots and AI’ (John Danaher), with AI-optimists generally hailing from outside philosophy.

In my presentation, I challenge the pessimistic sentiment within AI ethics by suggesting that it stems from a problematic negativity bias within the discipline. The problem, in a nutshell, is that AI ethicists have little choice but to come up with ethical concerns if they want to have a career. The incentives faced by AI ethicists must be assumed to lead to a systematic exaggeration of ethical problems with AI.

If this is correct, one takeaway is that AI is probably not as ethically problematic as the AI ethics community makes it out to be. We possess incriminating higher-order evidence regarding the community’s ability to correctly estimate how problematic AI is. It provides us with reason to assume that AI ethicists are ‘over-diagnosing’ ethical problems with AI, which entitles us to more positivity. Another lesson is that we should consider tweaking the incentives within the system to correct this dysfunction.

Is AI capable of creativity? This question is bound up with other challenging questions about the capacities of artificial systems. Human creativity typically involves some conscious experience of the creative project and some comprehension of the domain in which one is being creative. But are consciousness and comprehension necessary conditions of creativity? And, if so, what are the prospects of AI satisfying those conditions? I explore the role of consciousness and comprehension in the three stages of creativity – preparation, incubation and evaluation – and consider the challenges of attributing consciousness and comprehension to AI. I argue that although consciousness is not necessary for evaluation as such it is plausibly necessary for certain kinds of evaluation. Doubts about artificial consciousness then entail doubts about certain kinds of artificial creativity.

The integration of AI personal assistants into our daily lives promises to radically transform how we experience and represent ourselves. While technology’s ability to extend human agency has been widely discussed, AI assistants introduce a novel phenomenon: they can be simultaneously experienced as part of the self and as an other with whom we converse. Through an analysis of a near-future scenario involving an AI personal assistant, I show how these technologies can become transparently integrated into our perception and action, contribute to self-knowledge, and help us shape ourselves into who we want to be. At each stage, we encounter a peculiar duality: the AI assistant functions both as equipment that disappears from conscious awareness (becoming part of the pre-reflective sense of self) and as an interlocutor who provides an ‘insider’s outsider perspective’ on who we are. Rather than seeing this as undermining selfhood, I argue that this novel form of self-relation -— which I call the ‘self-as-other’ — may enhance our ability to know and shape ourselves. The paper thus contributes to debates about extended cognition and the impact of technology on human selfhood by identifying a novel way in which technology may transform self-experience: not through radical enhancement or replacement, but through the introduction of an other that is simultaneously experienced as self.

The application of AI in Medicine (AIM) is producing health practices more reliable, accurate and efficient than Traditional Medicine (TM) by assisting partly/totality of the medical decision-making, such as the use of deep learning in diagnostic imagery, designing treatment plans or preliminary diagnosis. Yet, most of these AI systems are pure “black-boxes”: the practitioner understands the inputs and outputs of the system but cannot have access to what happens “inside” it and cannot offer an explanation, creating an opaque process that culminates in a Trust Gap in two levels: (a) between patients and the medical experts; (b) between the medical expert and the medical process itself. This creates a “black-box medicine” since the practitioner ought to rely (epistemically) on these AI systems that are more accurate, fast and efficient but are not transparent (epistemically) and do not offer any kind of explanation. In this seminar, we aim to analyze a potential solution to address the Trust Gap in AI Medicine. We argue that a specific approach to Explainable AI (xAI) can succeed in reintroducing explanations into the discussion by focusing on how medical reasoning relies on social and abductive explanations and how AI can reproduce, potentially, this kind of abductive reasoning.