125293 research outputs found
Sort by
Will power‑seeking AGIs harm human society?
Full text linkMany have argued, based on the Instrumental Convergence Thesis, that Artificial General Intelligences (AGIs) will exhibit power-seeking behavior. Such behavior, they warn, could harm human society and pose existential threats—namely, the risk of human extinction or the permanent collapse of civilization. These arguments often rely on an implicit and underexamined assumption: that AGIs will develop world models—internal representations of world dynamics—that resemble those of humans. We challenge this assumption. We argue that once the anthropomorphic assumption—that AGIs’ world models will mirror our own—is rejected, it becomes unclear whether AGIs would pursue the types of power commonly emphasized in the literature, or any familiar types of power at all. This analysis casts doubt on the strength of existing arguments linking the Instrumental Convergence Thesis to existential threats. Moreover, it reveals a deeper layer of uncertainty. AGIs with non-human world models may identify novel or unanticipated types of power that fall outside existing taxonomies, thereby posing underappreciated risks. We further argue that world model alignment—an issue largely overlooked in comparison with value alignment—should be recognized as a core dimension of AI alignment. We conclude by outlining several open questions to inform and guide future research
Painting with Zombies: Neuroaesthetics and the Teleological Problem of Phenomenal Consciousness
Full text linkOne of the more pressing questions regarding phenomenal consciousness concerns its teleological function. For example, is there a purpose for having qualitative experiences when it seems at least conceivable to live as a zombie void of rich sensory experiences? In this paper, I will discuss recent findings in the emerging field of neuroaesthetics, which could form a novel framework for addressing what I will refer to as the teleological problem of phenomenal consciousness. While still in developmental stages, neuroaesthetics is moving beyond its early incarnations as a science that exclusively studies how the human brain responds to artworks, and into a science that studies how and, importantly, why we have sensory and qualitative experiences. I will ultimately argue that neuroaesthetics, when scaffolded by evolutionary principles, is poised to contribute empirically grounded responses to the teleological problem of phenomenal consciousness
Sophistry on Steroids? The Ethics, Epistemology and Politics of Persuasive AI
Full text linkThis paper examines the ethical, epistemological, and political implications of persuasive AI technologies. Recent research suggests that AI is roughly as persuasive as humans in many contexts. Should this concern us? I argue that, while some worries about persuasive AI may be overblown, we should be worried for a mix of ethical, epistemological and political reasons. Most centrally, we should be worried because persuasive AI may lead to a small number of powerful actors dominating what I call the “marketplace of arguments”—the set of arguments that provide the materials we use to discuss important moral, political and societal issues
Reflective Equilibrium
Full text linkHow can we figure out what’s right or wrong, if moral truths are neither self-evident nor something we can perceive? Very roughly, the method of reflective equilibrium (RE) says that we should begin moral inquiry from what we already confidently think, seeking to find a a match between our initial convictions and general principles that are well-supported by background theories, mutually adjusting both until we reach a coherent outlook in which our beliefs are in harmony (the equilibrium part) and we know why and how they support each other (the reflective part). It has been central to the self-understanding of normative ethics and other branches of philosophy in the last half a century. In this chapter, we examine the history of the idea of RE and introduce a schema for generating 256 variants. We explain why RE is subject to serious objections insofar as it purports to yield epistemic justification in virtue of achieving coherence. However, we also develop a new argument to the effect that RE is the best feasible method for us to achieve moral understanding and the ability to justify our judgments to others. It may thus be crucial for responsible moral inquiry, even if coherence among considered judgments and principles is neither sufficient nor necessary for justified moral belief
Sceptical hypotheses and subjective indistinguishability
Full text linkThe notion of subjective indistinguishability has long played a central role in explanations of the force of Cartesian sceptical hypotheses. I argue that sceptical hypotheses do not need to be subjectively indistinguishable to be compelling and I provide an alternative diagnosis of their force that explains why this is the case. My diagnosis focuses on the relation between one's experiences and third-personal accounts of the circumstances in which these experiences occur. This relation is characterized by a distinctive gap that leaves room for questions about the nature of one's circumstances, providing sceptical hypotheses with a foothold. I argue that this gap lends sceptical hypotheses their force and renders the stipulation of subjective indistinguishability unnecessary
More scepticism about epistemic blame
Full text linkScepticism about epistemic blame maintains that there is no distinctly epistemic form of blame. Cameron Boult (2024a) challenges this view on the grounds that a particular theory of epistemic blame—the relationship modification account—can be defended against the sceptic. His defence includes two central claims. First, Boult proposes a pluralism in our repertoire of epistemic accountability practices that makes space for both epistemic blame and epistemic evaluation. Second, he develops a more ’sceptic-resistant’ way of articulating the phenomenon that proponents of epistemic blame believe a theory of epistemic blame is needed to explain. In my view, neither of these arguments should convince us. I argue that the epistemic blame sceptic can meet both challenges. Although my arguments largely focus on Boult’s defence of his preferred account of epistemic blame, I also set out two general problems that faces any theory of epistemic blame. The upshot is that there remain strong grounds for scepticism about epistemic blame
Logopsychism: The Meaning of Consciousness
No full textLogopsychism is introduced as an ontologically neutral theory of consciousness that posits its primary function as the ascription of subjective meaning. It shifts the discussion from what consciousness is to what consciousness is for. Its framework is built upon four axioms: (1) the universe is a source of infinite meaning potential; (2) consciousness is the structure through which meaning is ascribed to all events; (3) subjectivity interprets this meaning into potential lived experience; and (4) consciousness either accepts this interpretation or recursively layers new meaning until an acceptable condition is reached. Meaning can be layered and mapped on a three-dimensional plane along the axes of depth, breadth, and time. Logopsychism is presented as a supplemental theory that can be integrated into any account that recognizes meaning and subjectivity as integral to first-person experience