AI and Fairness:

Toward a Philosophical Taxonomy of Bias in Large Language Models: A Survey

Jack Jacobs and Tania Aiyar authored the philosophical section of an interdisciplinary survey paper with computer scientists from Rutgers and UT Dallas. Their contribution compares philosophical and computational conceptions of bias, highlighting key points of convergence and divergence.

https://www.techrxiv.org/doi/full/10.36227/techrxiv.177273600.04458982/v1

AI and Griefbots:

From Familiarity to Transformation: Rethinking Griefbots in the Grieving 

(Jack Jacobs, Alex Lau, Amani El-Sheikh, Alexis Woods, Qiannan Li)

Abstract: Griefbots are increasingly promoted as tools for bereavement. By simulating the speech, personality, or presence of the dead, they appear to offer comfort, familiarity, and a way of sustaining continuing bonds after loss. In this paper, we argue that their prudential value is more limited than defenders of the familiarity account suggest. Our objection is not to continuing bonds as such, nor to technological mediation in grief more generally, but to the mimetic and quasi-reciprocal way griefbots operationalize those bonds. Grief, we contend, is not merely a backward-looking response to what has been lost. It is also a transformative process through which the bereaved confront an altered present and reorganize possible futures. For that reason, grief can make available distinctive prudential goods, including narrative reconstitution, biographical self-knowledge, value disclosure, and practical reorientation. Griefbots are normatively risky, we argue, because they can impede access to these goods by reinforcing loss-oriented coping and simulating ongoing reciprocity with the deceased. At the same time, positive testimony from griefbot users should not be dismissed: such reports provide genuine evidence of comfort, companionship, and emotional support. But immediate consolation does not by itself settle the broader prudential question. We therefore propose a different design orientation for grief-related AI: transformative bots. Rather than simulating the dead, such systems would function as non-mimetic companions for remembrance, reflection, and restoration. If AI is to play a constructive role in bereavement, we argue, it should help the living reinterpret loss and learn how to live forward under conditions of absence.

AI and Interesting Life:

Beyond Comfort: Interesting Experiences and the Measurement of Well-Being in AI Care Companions

(Jack Jacobs, Alex Lau, Daezhon Gilbarg, Hanxuan Zhang, Qiannan Li)

Abstract:

This paper argues that AI care companions for older adults should be evaluated by a richer conception of well-being than comfort, companionship, mood improvement, or loneliness reduction alone. Drawing on Lorraine Besser’s account of interesting experiences, we argue that interesting experiences constitute a distinct experiential good: they involve robust cognitive engagement, can be valuable even when they are not simply pleasant, and often open further possibilities for curiosity, reflection, and exploration. Using ElliQ as a philosophical case study, we show how a welfare-oriented AI companion may provide genuine benefits while still reflecting a narrow conception of well-being. ElliQ’s public-facing materials and evaluation framework foreground reassurance, positivity, companionship, and broad mental activity, but they do not appear to treat interesting experiences as a distinct welfare target. This creates a measurement gap: if AI companions are assessed mainly by whether users feel accompanied, cheered, or generally engaged, then important non-hedonic dimensions of well-being may remain invisible. In response, we propose three design principles—novelty, challenge, and relational memory—that can help AI companions make interesting experiences more reliably available. The aim is not to replace comfort with challenge, but to ensure that AI care technologies help keep the world open.

AI and Loneliness: 

A project arguing that AI cannot meaningfully alleviate loneliness on the grounds that it fails to fulfill the moral obligations between persons required to resolve loneliness.

AI and Alignment:

Project 1: Moral Advising by AI Systems

Project 2: Persuasive AI in Medical Decision-Making

AI and Moral Responsibility:

Project 1: Moral Responsibility Gaps in Driverless Car Systems