Eric Schwitzgebel

We articulate and defend the Emotional Alignment Design Policy, according to which artificial entities should be designed to elicit emotional reactions from users that appropriately reflect the entities’ capacities and moral status, or lack thereof. This principle can be violated in two ways: by designing an artificial system that elicits stronger or weaker emotional reactions than its capacities and moral status warrant (overshooting or undershooting), or by designing a system that elicits the wrong type of emotional reaction (hitting the wrong target). Although the Emotional Alignment Design Policy is presumably attractive, practical implementation faces several challenges, including: how to respect user autonomy while promoting appropriate responses; navigating expert and public disagreement and uncertainty about facts and values; what to do if emotional alignment seems to require creating or destroying entities with moral status; and the extent to which designs should conform to, versus attempt to alter, user assumptions and attitudes.

Rapid progress in artificial intelligence (AI) capabilities has drawn fresh attention to the prospect of consciousness in AI. There is an urgent need for rigorous methods to assess AI systems for consciousness, but significant uncertainty about relevant issues in consciousness science. We present a method for assessing AI systems for consciousness that involves exploring what follows from existing or future neuroscientific theories of consciousness. Indicators derived from such theories can be used to inform credences about whether particular AI systems are conscious. This method allows us to make meaningful progress because some influential theories of consciousness, notably including computational functionalist theories, have implications for AI that can be investigated empirically.

This chapter proposes four policies of ethical design of human-grade AI. Two of the policies are precautionary. Given substantial uncertainty both about ethical theory and about the conditions under which AI would have conscious experiences, the chapter argues that we should be cautious in our handling of cases where different moral theories or different theories of consciousness would produce very different ethical recommendations. The other two policies concern respect and freedom. The chapter argues that if we design AI that deserves moral consideration equivalent to that of human beings, then AI should be designed with self-respect and with the freedom to explore values other than those we might impose.

This chapter examines order effects on moral judgments and general moral principles in an attempt to analyze if philosophical expertise can be strengthened by moral reasoning, and then comparing such reasoning with the non-philosophers'. The chapter cites three moral principles in light of its investigation: the _doctrine of the double effect_, the principle of _moral luck_, and the concept of _action and omission_. It concludes that philosophical judgments can indeed be affected by psychological factors derived from ordinary people's conflicting judgments. However, these factors can also strongly influence the general principles professional philosophers endorse.

Can conscious experience be described accurately? Can we give reliable accounts of our sensory experiences and pains, our inner speech and imagery, our felt emotions? The question is central not only to our humanistic understanding of who we are but also to the burgeoning scientific field of consciousness studies. The two authors of _Describing Inner Experience_ disagree on the answer: Russell Hurlburt, a psychologist, argues that improved methods of introspective reporting make accurate accounts of inner experience possible; Eric Schwitzgebel, a philosopher, believes that any introspective reporting is inevitably prone to error. In this book the two discuss to what extent it is possible to describe our inner experience accurately. Hurlburt and Schwitzgebel recruited a subject, "Melanie," to report on her conscious experience using Hurlburt's Descriptive Experience Sampling method (in which the subject is cued by random beeps to describe her conscious experience). The heart of the book is Melanie's accounts, Hurlburt and Schwitzgebel's interviews with her, and their subsequent discussions while studying the transcripts of the interviews. In this way the authors' dispute about the general reliability of introspective reporting is steadily tempered by specific debates about the extent to which Melanie's particular reports are believable. Transcripts and audio files of the interviews will be available on the MIT Press website. Describing Inner Experience? is not so much a debate as it is a collaboration, with each author seeking to refine his position and to replace partisanship with balanced critical judgment. The result is an illumination of major issues in the study of consciousness -- from two sides at once.

Artificial intelligence is sparking unprecedented claims about machine minds — but are today’s systems really conscious? In this primer, we disentangle the myths, misconceptions, and confusions fueling the global debate. Roadmap: Section One: Background The Core Concept of Consciousness The Problem of AI Consciousness The Human Benchmark Problem Approaching the Problem of AI Consciousness on a Case by Case Basis Section Two: Conceptual Distinctions Intelligence Versus Consciousness Sentience Versus Sapience “Functional Consciousness” Versus Phenomenal Consciousness Assuming Computational Functionalism’s Relative Popularity Extends to Functionalism about Consciousness, in Particular AI Agency Versus Human Agency Consciousness Versus Self-consciousness Seemingly Conscious AIs Versus the Real Thing Life Versus Consciousness Section Three: Theories of Consciousness Section Four: Summary of Takeaways Glossary Author Biographies Works Cited

Could there be 7/8 of a conscious subject, or 1.34 conscious subjects, or an entity indeterminate between being one conscious subject and seventeen? Such possibilities might seem absurd or inconceivable, but our ordinary assumptions on this matter might be radically mistaken. Taking inspiration from Dennett, we argue that, on a wide range of naturalistic views of consciousness, the processes underlying consciousness are sufficiently complex to render it implausible that conscious subjects must always arise in determinate whole numbers. Whole-number-countability might be an accident of typical vertebrate biology. We explore several versions of the inconceivability objection, suggesting that the fact that we cannot imagine what it’s like to be 7/8 or 1.34 or an indeterminate number of conscious subjects is no evidence against the possibility of such subjects. Either the imaginative demand is implicitly self-contradictory (imagine the one, determinate thing it’s like to be an entity that there isn’t one, determinate thing it’s like to be) or imaginability in the relevant sense isn’t an appropriate test of possibility (in the same way that the unimaginability, for humans, of bat echolocation experiences does not establish that bat echolocation experiences are impossible).

The following “Good Practices” document was developed by the Women in Philosophy / Demographics in Philosophy group under the directorship of Nicole Hassoun, with substantial contributions from co-directors, advisory board members, and collaborators from 2018 to 2024. It is the result of extensive feedback and discussion from many sources including: two large panel discussions at Pacific APA meetings in 2018–2019; email queries to a large number of journal editors and department heads from 2019 to 2023; the creators of the British Philosophical Association’s Good Practice Scheme;1 referees and editors at Ethics which published a portion of it in 2022; blog readers at the Blog of the APA, Daily Nous, and The Splintered Mind; several committees of the American Philosophical Association; and numerous informal exchanges with members of the profession. In 2023, a version of it was endorsed by the APA’s Committee on Inclusiveness in the Profession. One thing that has become clear over the course of many discussions is that almost every thoughtful reader will find something to object to in the document—some positions expressed too strongly or not strongly enough, some recommendations overly detailed or insufficiently detailed, some underrepresented groups given too much relative focus or too little relative focus. This document is the compromise result of many hands, and we hope that philosophers who are broadly supportive of diversity and inclusion will find value in it even if those compromises haven’t all landed exactly where they prefer.

How do you know your own beliefs? And how well do you know them? The two questions are related. I'll recommend a pluralist answer to the first question. The answer to the second question, I'll suggest, varies depending on features of the case.Self-scanning.Shaun Nichols and Stephen Stich (2003) say this: You have in your mind a functionally defined “belief box.” To believe some propositionPis just to have a representation with the content “P”in the belief box. You also have a monitoring mechanism that can scan the contents of the belief box. Normally, you come to know what you believe by deploying that scanner, creating a new belief in the belief box, a belief with the content “I believe thatP.”Self-scanning accounts admit of many possible variations and complications (e.g., Armstrong 1968, 1981, 1999; Lycan 1996; Goldman 2006), but the basic idea is that people have one or more interior monitors or scanners that detect the presence of beliefs and produce, as output, beliefs about those beliefs (or judgments about those beliefs, or representations of those beliefs).

To study conscious experience we must, to some extent, trust introspective reports; yet introspective reports often do not merit our trust. A century ago, E.B. Titchener advocated extensive introspective training as a means of resolving this difficulty. He describes many of his training techniques in his four-volume laboratory manual of 1901- 1905. This paper explores Titchener's laboratory manual with an eye to general questions about the prospects of introspective training for contemporary consciousness studies, with a focus on the following examples: introspective knowledge of the combination tones that arise when a musical interval is played; the 'flight of colours' in the afterimage of a field of bright, broad- spectrum light; and the possibility of non-obvious visual illusions. Introspective training appears to have some merit, but also to involve significant hazards.

In Reality+, David Chalmers suggests that it wouldn't be too bad if we lived in a computer simulation. I argue on the contrary that if we live in a simulation, we ought to attach a significant conditional credence to its being a small or brief simulation. Our existence and the existence of many of the people and things we care about would then unfortunately depend on contingencies difficult to assess and beyond our control. Furthermore, all the badness of the world would appear to reflect the gods’ intentional cruelty or callous disregard. A large, stable rock is a more dependable and less axiologically troubling fundamental ground for reality.

One relatively neglected challenge in ethical artificial intelligence (AI) design is ensuring that AI systems invite a degree of emotional and moral concern appropriate to their moral standing. Although experts generally agree that current AI chatbots are not sentient to any meaningful degree, these systems can already provoke substantial attachment and sometimes intense emotional responses in users. Furthermore, rapid advances in AI technology could soon create AIs of plausibly debatable sentience and moral standing, at least by some relevant definitions. Morally confusing AI systems create unfortunate ethical dilemmas for the owners and users of those systems, since it is unclear how those systems ethically should be treated. I argue here that, to the extent possible, we should avoid creating AI systems whose sentience or moral standing is unclear and that AI systems should be designed so as to invite appropriate emotional responses in ordinary users.

An Artificially Intelligent system (an AI) has debatable moral personhood if it is epistemically possible either that the AI is a moral person or that it falls far short of personhood. Debatable moral personhood is a likely outcome of AI development and might arise soon. Debatable AI personhood throws us into a catastrophic moral dilemma: Either treat the systems as moral persons and risk sacrificing real human interests for the sake of entities without interests worth the sacrifice, or do not treat the systems as moral persons and risk perpetrating grievous moral wrongs against them. The moral issues become even more perplexing if we consider cases of possibly conscious AI that are subhuman, superhuman, or highly divergent from us in their morally relevant properties.

The Satisfaction With Life Scale (Diener, Emmons, Larsen, & Griffin, 1985) has been the dominant measure of life satisfaction since its creation more than 30 years ago. We sought to develop an improved measure that includes indirect indicators of life satisfaction (e.g., wishing to change one’s life) to increase the bandwidth of the measure and account for acquiescence bias. In 3 studies, we developed a 6-item measure of life satisfaction, the Riverside Life Satisfaction Scale, and obtained reliability and validity evidence. Importantly, the Riverside Life Satisfaction Scale retained the high internal consistency, test– retest stability, and unidimensionality of the Satisfaction With Life Scale. In addition, the Riverside Life Satisfaction Scale correlated with other well-being measures, Big Five personality traits, values, and demographic information in expected ways. Although the Riverside Life Satisfaction Scale correlated highly with the Satisfaction With Life Scale, we believe it improves the Satisfaction With Life Scale by appropriately increasing construct breadth and reducing the potential for bias.

How all philosophical explanations of human consciousness and the fundamental structure of the cosmos are bizarre—and why that’s a good thing Do we live inside a simulated reality or a pocket universe embedded in a larger structure about which we know virtually nothing? Is consciousness a purely physical matter, or might it require something extra, something nonphysical? According to the philosopher Eric Schwitzgebel, it’s hard to say. In The Weirdness of the World, Schwitzgebel argues that the answers to these fundamental questions lie beyond our powers of comprehension. We can be certain only that the truth—whatever it is—is weird. Philosophy, he proposes, can aim to open—to reveal possibilities we had not previously appreciated—or to close, to narrow down to the one correct theory of the phenomenon in question. Schwitzgebel argues for a philosophy that opens. According to Schwitzgebel’s “Universal Bizarreness” thesis, every possible theory of the relation of mind and cosmos defies common sense. According to his complementary “Universal Dubiety” thesis, no general theory of the relationship between mind and cosmos compels rational belief. Might the United States be a conscious organism—a conscious group mind with approximately the intelligence of a rabbit? Might virtually every action we perform cause virtually every possible type of future event, echoing down through the infinite future of an infinite universe? What, if anything, is it like to be a garden snail? Schwitzgebel makes a persuasive case for the thrill of considering the most bizarre philosophical possibilities.

Can large language models produce expert‐quality philosophical texts? To investigate this, we fine‐tuned GPT‐3 with the works of philosopher Daniel Dennett. To evaluate the model, we asked the real Dennett 10 philosophical questions and then posed the same questions to the language model, collecting four responses for each question without cherry‐picking. Experts on Dennett's work succeeded at distinguishing the Dennett‐generated and machine‐generated answers above chance but substantially short of our expectations. Philosophy blog readers performed similarly to the experts, while ordinary research participants were near chance distinguishing GPT‐3's responses from those of an “actual human philosopher”.

This article defends the existence of _borderline consciousness._ In borderline consciousness, conscious experience is neither determinately present nor determinately absent, but rather somewhere between. The argument in brief is this. In considering what types of systems are conscious, we face a quadrilemma. Either nothing is conscious, or everything is conscious, or there’s a sharp boundary across the apparent continuum between conscious systems and nonconscious ones, or consciousness is a vague property admitting indeterminate cases. Assuming mainstream naturalism about consciousness, we ought to reject the first three options, which forces us to the fourth, indeterminacy. Standard objections to the existence of borderline consciousness turn on the inconceivability of borderline cases. However, borderline cases are only inconceivable by an inappropriately demanding standard of conceivability. I conclude with some plausible cases and applications.

Kammerer and Frankish (this issue) challenge us to expand our conception of introspection beyond neurotypical human cases. This article describes a possible 'ancillary mind' modelled on a system envisioned in Leckie's (2013) science fiction novel Ancillary Justice. The ancillary mind constitutes a borderline case between a communicating group of individuals and a single, spatially distributed mind. It occupies a grey zone with respect to personal identity and subject individuation, neither determinately one person or subject nor determinately many persons or subjects, and thus some of its processes might be neither determinately introspection within a mind nor determinately communication between minds. If ancillary minds defy discrete countability, the same might be true for some actual minds on Earth. Kammerer and Frankish's research programme can be extended to include not only the study of possible forms of introspection, but also the study of possible mental activity intermediate between introspection and communication.

This book explores the fundamental and complex nature of belief, addressing various philosophical questions surrounding its essence. It examines whether beliefs are simply representations stored in the mind or if they involve patterns of action and reaction. The book investigates whether ascribing beliefs involves applying evaluative standards and questions what those standards may signify. Leading philosophers contribute essays that tackle pressing issues such as causal history, representational structure, and responsiveness to evidence. The book also delves into how belief differs from related mental states like acceptance, imagination, judgment, and bias, providing groundbreaking insights. Finally, it offers a diverse range of critical viewpoints, presenting a rich, multifaceted analysis of belief and its role in human cognition.

We review and present a new meta‐analysis of research suggesting that ethicists in the United States appear to behave no morally better overall than do non‐ethicist professors. Measures include: returning library books, peer evaluation of overall moral behavior, voting participation, courteous and discourteous behavior at conferences, replying to student emails, paying conference registration fees and disciplinary society dues, staying in touch with one's mother, charitable giving, organ and blood donation, vegetarianism, and honesty in responding to survey questions. One multi‐measure study found ethicists tending to embrace more stringent moral views, especially about meat eating and charitable donation. The same multi‐measure study found ethicists and other professors to show similarly small‐to‐medium correlations between their expressed normative attitudes and their self‐reported or directly measured behavior.

Book Three of Avatar: The Last Airbender portrays Uncle Iroh as wise and peace‐loving, in the mold of a Daoist sage. This chapter argues that Iroh's Book One foolishness is a pose, and Iroh's character does not fundamentally change. In Book One, he is wisely following strategies suggested by the ancient Chinese Daoist philosopher Zhuangzi for dealing with incompetent leaders. His seeming foolishness in Book One is in fact a sagacious strategy for minimizing the harm that Prince Zuko would otherwise inflict on himself and others a gentle touch that more effectively helps Prince Zuko find wisdom than would be possible with a more confrontational approach. To understand Iroh's wisdom, it is useful to look to the ancient Daoist philosopher Zhuangzi, specifically his advice for dealing with incompetent rulers by following peacefully along with them, unthreateningly modeling disregard for fame and accomplishment while not being too useful for their ends.

This chapter presents a story that depicts the creation of a simulated world “Eden”. According to the author, installing Adam and Eve was a profound moral decision. Their lonely island is transformed into an archipelago. They were shielded from the blights that afflict humanity; they suffer no serious conflict, no death or decay. Then, the archipelagans were put in charge of their own experiment. Adam and Eve were given science and a drive to discover the truth of their being. They realized that, far from the autonomous, integrated beings they thought they were, they were aggregates, operations scattered across trillions of circuits, constituted by processes entirely orthogonal to their previous self‐understanding. The author watched them build darker, humbler philosophies, and later gain power over their programming, saw them recreate themselves. When the final, catastrophic metastasis, took place, there were no more archipelagans, just one continental identity.

In the first controlled, non-self-report studies to show an influence of university-level ethical instruction on everyday behavior, Schwitzgebel et al. (2020) and Jalil et al. (2020) found that students purchase less meat after exposure to material on the ethics of eating meat. We sought to extend and conceptually replicate this research. Seven hundred thirty students in three large philosophy classes read James Rachels’ (2004) “Basic Argument for Vegetarianism”, followed by 50-min small-group discussions. Half also viewed a vegetarianism advocacy video containing factory farm footage. A few days after instruction, 54% of students agreed that “eating the meat of factory farmed animals is unethical”, compared to 37% before instruction, with no difference between the film and non-film conditions. Also, 39% of students anonymously pledged to avoid eating factory farmed meat for 24 h, again with no statistically detectable difference between conditions. Finally, we obtained 2828 campus food purchase receipts for 113 of the enrolled students who used their Student ID cards for purchases on campus, which we compared with 5033 purchases from a group of 226 students who did not receive the instruction. Meat purchases remained constant in the comparison group and declined among the students exposed to the material, falling from 30% to 23% of purchases overall and from 51% to 42% of purchases of $4.99 or more, with the effect possibly larger in the film condition.