Fabio Paglieri

This paper discusses some societal implications of the most recent and publicly discussed application of advanced machine learning techniques: generative AI models, such as ChatGPT (text generation) and DALL-E (text-to-image generation). The aim is to shift attention away from conceptual disputes, e.g. regarding their level of intelligence and similarities/differences with human performance, to focus instead on practical problems, pertaining the impact that these technologies might have (and already have) on human societies. After a preliminary clarification of how generative AI works (Sect. 1), the paper discusses what kind of transparency ought to be required for such technologies and for the business model behind their commercial exploitation (Sect. 2), what is the role of user-generated data in determining their performance and how it should inform the redistribution of the resulting benefits (Sect. 3), the best way of integrating generative AI systems in the creative job market and how to properly negotiate their role in it (Sect. 4), and what kind of “cognitive extension” offered by these technologies we ought to embrace, and what type we should instead resist and monitor (Sect. 5). The last part of the paper summarizes the main conclusions of this analysis, also marking its distance from other, more apocalyptic approaches to the dangers of AI for human society.

As state-of-art technology can create artificial images that are indistinguishable from real ones, it is urgent to understand whether believing that a picture is real or not has some import over affective phenomena such as sexual arousal. Thus, in two pre-registered online studies, we tested whether 60 images depicting models in underwear elicited higher self-reported sexual arousal when believed to be (N = 57) or presented as (N = 108) real photographs as opposed to artificially generated. In both cases, Realness correlated with significantly higher scores on self-reported sexual arousal. Consistently with the literature on downregulation of emotional response to fictional works, our result indicates that sexual images that are perceived to be fake are less arousing than those believed to portray real people.

Enthymemes are traditionally defined as arguments in which some elements are left unstated. It is an empirical fact that enthymemes are both enormously frequent and appropriately understood in everyday argumentation. Why is it so? We outline an answer that dispenses with the so called "principle of charity", which is the standard notion underlying most works on enthymemes. In contrast, we suggest that a different force drives enthymematic argumentation—namely, parsimony, i.e. the tendency to optimize resource consumption, in light of the agent's goals. On this view, the frequent use of enthymemes does not indicate sub-optimal performance of arguers, requiring appeals to charity for their redemption. On the contrary, it is seen as a highly adaptive argumentation strategy, given the need of everyday reasoners to optimize their cognitive resources. Considerations of parsimony also affect enthymeme reconstruction, i.e. the process by which the interpreter makes sense of the speaker's enthymemes. Far from being driven by any pro-social cooperative instinct, interpretative efforts are aimed at extracting valuable information at reasonable costs from available sources. Thus, there is a tension between parsimony and charity, insofar as the former is a non-social constraint for self-regulation of one's behaviour, whereas the latter implies a pro-social attitude. We will argue that some versions of charity are untenable for enthymeme interpretation, while others are compatible with the view defended here, but still require parsimony to expose the ultimate reasons upon which a presumption of fair treatment in enthymeme reconstruction is founded

Traditionally, an enthymeme is an incomplete argument, made so by the absence of one or more of its constituent statements. An enthymeme resolution strategy is a set of procedures for finding those missing elements, thus reconstructing the enthymemes and restoring its meaning. It is widely held that a condition on the adequacy of such procedures is that statements restored to an enthymeme produce an argument that is good in some given respect in relation to which the enthymeme itself is bad. In previous work, we emphasized the role of parsimony in enthymeme resolution strategies and concomitantly downplayed the role of charity . In the present paper, we take the analysis of enthymemes a step further. We will propose that if the pragmatic features that attend the phenomenon of enthymematic communication are duly heeded, the very idea of reconstructing enthymemes loses much of its rationale, and their interpretation comes to be conceived in a new light

Philosophers of science have given up on the quest for a silver bullet to put an end to all pseudoscience, as such a neat formal criterion to separate good science from its contenders has proven elusive. In the literature on critical thinking and in some philosophical quarters, however, this search for silver bullets lives on in the taxonomies of fallacies. The attractive idea is to have a handy list of abstract definitions or argumentation schemes, on the basis of which one can identify bad or invalid types of reasoning, abstracting away from the specific content and dialectical context. Such shortcuts for debunking arguments are tempting, but alas, the promise is hardly if ever fulfilled. Different strands of research on the pragmatics of argumentation, probabilistic reasoning and ecological rationality have shown that almost every known type of fallacy is a close neighbor to sound inferences or acceptable moves in a debate. Nonetheless, the kernel idea of a fallacy as an erroneous type of argument is still retained by most authors. We outline a destructive dilemma we refer to as the Fallacy Fork: on the one hand, if fallacies are construed as demonstrably invalid form of reasoning, then they have very limited applicability in real life. On the other hand, if our definitions of fallacies are sophisticated enough to capture real-life complexities, they can no longer be held up as an effective tool for discriminating good and bad forms of reasoning. As we bring our schematic “fallacies” in touch with reality, we seem to lose grip on normative questions. Even approaches that do not rely on argumentation schemes to identify fallacies fail to escape the Fallacy Fork, and run up against their own version of it

Elefanti nani, inesistenti regni medievali, parodie politiche scambiate per vere, panzane virali e dibattiti privi di senso sui social media. Le bufale un tempo erano oggetto di curiosità, bizzarri orpelli della credulità umana di cui discutere fra il serio e il faceto. Oggi causano allarme sociale, come ci dimostra la cronaca recente: la baraonda digitale prodotta dal diffondersi di un'epidemia può minare alla radice i tentativi di combatterla, o al contrario facilitare una risposta collettiva sensata ed efficace. Perché la disinformazione online è soprattutto il sintomo di cambiamenti radicali nelle nuove tecnologie di comunicazione. Convivere felicemente con tutto questo è possibile, e persino necessario. Per farlo, bisogna considerare le bufale non spazzatura di cui sbarazzarsi, ma piuttosto fantastici laboratori su cui affinare le nostre competenze. Una proposta originale per orientarsi meglio nella nuova ecologia dell'informazione--Page 2 of cover.

The central focus of this research is the fast and crucial impact of the COVID-19 pandemic, and its exceptionally serious consequences in terms of healthcare, state intervention and impositions, radical changes in people’s life, on a crucial psychological, relational, and political construct: trust. In this survey, addressed to 4260 Italian citizens, we tried to analyze and measure such impact, focusing on various aspects of trust. This attention to multiple dimensions of trust constitutes the key conceptual advantage of this research, since trust is a complex and layered construct, with its own internal dynamics. In particular, the analysis focuses on how citizens attribute trust to Public Authorities, in relation to the management of the health crisis: with regard to the measures and guidelines adopted, the purposes pursued, the motivations that determine them, their capacity for involvement and their effectiveness for the containment of the virus itself. A pandemic creates a bilateral need for trust, both in Public Authorities (they have to rely on citizens’ compliance, and must try to promote and maintain their trust in order to be effective), and in citizens, since they need to feel that somebody can do something, can (has the power) to protect them, to act at the needed collective level. We are interested to explore how this need for trust affects the attributional process, regarding both attitudes and the corresponding decisions and actions. The most striking result of this survey is the very high level of institutional trust expressed by respondents: 75% of them trust Italian public authorities to be able to deal with the COVID-19 emergency. This is in sharp contrast with the relatively low levels of institutional trust characteristic of Italy, both historically and in recent surveys. Moreover, the survey allowed to discriminate several potential predictors for trust, thus emphasizing factors that, during this crisis, are exhibiting an anomalous impact on trust.

Why are enthymemes so frequent? Are we dumb arguers, smart rhetoricians, or parsimonious reasoners? This paper investigates systematic use of enthymemes, criticizing the application of the principle of charity to their interpretation. In contrast, I propose to analyze enthymematic argumentation in terms of parsimony, i.e. as a manifestation of the rational tendency to economize over scant resources. Consequences of this view on the current debate on enthymemes and on their rational reconstruction are discussed

From a decision theoretic perspective, arguments stem from decisions made by arguers. Despite some promising results, this approach remains underdeveloped in argumentation theories, mostly because it is assumed to be merely descriptive. This assumption is mistaken: considering arguments as the product of decisions brings into play various normative models of rational choice. The challenge is rather to reconcile strategic rationality with other normative constraints relevant for argumentation, such as inferential validity and dialectical appropriateness.

Illusionism is a prominent hypothesis about action control, according to which acts that we consider voluntary are nevertheless caused by unconscious brain events, and thus our subjective experience of consciously willing them is ultimately illusory. Illusionism can be understood as either an ontological thesis or a phenomenological claim, but both versions are vulnerable to a line of attack based on the role of long-term planning in action control. According to this objection, the evidence upon which illusionism rests is confined to short-term intentions, so it is not sufficient to justify broader conclusions on the causal inefficacy of conscious will. In this essay we reconstruct the logic of this objection against illusionism, clarify why surveying folk intuitions on conscious distal intentions is essential to the debate, and present a study in which the role of conscious planning in intentionality judgment is clearly revealed. We also present other relevant findi...

In this brief commentary of Kamila Debowska-Kozlowska’s insightful analysis of persuasive outcomes (Processing topics from the Beneficial Cognitive Model in partially and over-successful persuasion dialogues. Argumentation, 2014), I articulate some suggestions for future development of her ideas. My main claim is that, while instances of partially and over-successful persuasion are indeed worthy of further theoretical inquiry, the topical analysis proposed by Debowska-Kozlowska may benefit from integration with other approaches

In spite of significant research efforts, argument technologies do not seem poised to scale up as much as most commentators would hope or even predict. In this paper, I discuss what obstacles bar the way to more widespread success of argument technologies and venture some suggestions on how to circumvent such difficulties: doing so will require a significant shift in how this research area is typically understood and practiced. I begin by exploring a much broader yet closely related question: To what extent are people natively good at arguing? This issue has always been central to philosophical reflection and it has become even more urgent nowadays, with the explosion of persuasive technologies and unprecedented opportunities for large-scale social influence. The answer hinges on what aspect of argumentation is taken under consideration: evidence suggests that people are relatively bad at analyzing the structure of arguments, especially when these are presented out of context and in abstract terms; in contrast, data show that even laymen tend to excel in the interactive practice of argumentation, in particular when motivation is high and something significant is at stake. Unfortunately, current argument technologies are more closely tailored to the former type of activity than to the latter, which is the main reason behind their relative lack of success with the general public. Changing this state of affair will require a commitment to ecological argument technologies: that is, technologies designed to support real-time, engaging and meaningful argumentative interactions performed by laypeople in their ordinary life, instead of catering to the highly specific needs of a minority of niche users.