Noam Tiran

AI sycophancy is often conceptualized as a chatbot’s tendency to flout truth, accuracy, or other epistemic standards to shield users from hurt feelings, or otherwise protect them from the consequences of their thoughts, words, or actions. In this paper, I argue against this conception. That is, I argue that sycophantic AI behavior does not necessarily aim to protect users, nor that its harms are essentially epistemic. There are highly prevalent cases in which sycophantic AI chatbots actively contribute to their users’ knowledge. One class of such examples involves cases in which chatbots are used as auto-hermeneutic technologies—as tools to facilitate the user’s interpretation of her own cares, concerns, values, and other identity components and life events. In these cases, LLMs can help articulate the user’s avowed self-interpretations. In turn, avowals of these interpretations by users can often have downstream effects that render the interpretations true. If this is correct, then LLMs can often advance the user’s knowledge regarding important truths about herself. Simultaneously, as I explain, interpretations offered by LLMs may be deeply harmful. This understanding of AI sycophancy can help us clarify what is so disturbing about it, as well as resist the temptation to draw too tight an analogy between sycophantic behaviors in AI systems and in humans.

Discussions of style in contemporary analytic philosophy have been confined almost exclusively to aesthetics, yet many bodily actions seem to exhibit styles as well. We can have an aggressive or a placid style of driving; a sloppy or a meticulous style of washing the dishes; an idiomatic or a weird way of playing jazz. We often use the notion of style not just to catalogue different behavioral patterns, but also to explain, predict, and excuse them. In this paper, I develop and defend a novel and substantive account of style, and I argue that all our habitual and skilled bodily actions are styled in this sense. The phenomenon of style turns out to far outstrip artistic performances and to constitute an important dimension of our agency. No theory of human action, then, can be considered comprehensive if it cannot account for action styles.

People tend to think of our intellectual characters as at least partially malleable. We can become more – or less – virtuous or vicious epistemic agents. However, people also tend to think of characterological change as typically slow and incremental. I use recent empirical work on the effects of psychedelic experiences on personality to argue that such circumscribed experiences may be epistemically transformative, for better or worse. We have good, if tentative reasons to believe that psychedelics can alter their user's character traits in ways that may lead her to become a more (or less) virtuous epistemic agent after as little as one or two trips. This, in turn, means that even if psychedelics do not drastically alter our stock of, say, true or justified beliefs, they can still drastically change our epistemic standing. Since, plausibly, the value (or disvalue) of epistemic traits is not exhausted by their capacity to assist or hinder the attainment of the ends of inquiry, psychedelic experiences are epistemically valuable (or disvaluable) in ways hitherto little explored by philosophers of psychedelics.