Occasional Paper
In a new working paper, Crocodile Tears: Can the Ethical-Moral Intelligence of AI Models Be Trusted?, Sarah Hubbard, Associate Director for Technology & Democracy, David Kidd, an Allen Lab member, and Andrei Stupu, a former Allen Lab fellow, introduce a framework for evaluating the ethical-moral intelligence of AI models across dimensions of moral expertise, sensitivity, coherence, and transparency.
This paper was also published in AI and Ethics, you can find it here.
Abstract
As AI becomes increasingly embedded into every aspect of our lives, there is evidence that people are turning to these systems for guidance on complex issues and moral dilemmas. Whether or not one agrees that people should do so, the fact that they are necessitates a clearer understanding of the moral reasoning of these systems. To address this gap, Crocodile Tears: Can the Ethical-Moral Intelligence of AI Be Trusted? introduces an ethical-moral intelligence (EMI) framework for evaluating AI models across dimensions of moral expertise, sensitivity, coherence, and transparency. We present findings from a pre-registered experiment testing the moral sensitivity in four AI models (Claude, GPT, Llama, and DeepSeek) using ethically challenging scenarios. While models demonstrate moral sensitivity to ethical dilemmas in ways that closely mimic human responses, they exhibit greater certainty than humans when choosing between conflicting sacred values, despite recognizing such tragic trade-offs as difficult. This discrepancy between reported difficulty and decisiveness raises important questions about their coherence and transparency, undermining trustworthiness. The research reveals a critical need for more comprehensive ethical evaluation of AI systems. We discuss the implications of these specific findings, how psychological methods might be applied to understand the ethical-moral intelligence of AI models, and outline recommendations for developing more ethically aware AI that augments human moral reasoning.
Sarah Hubbard is the Associate Director for Technology & Democracy at the Ash Center’s Allen Lab for Democracy Renovation and was previously a Technology & Public Purpose Fellow at the Belfer Center.
David Kidd is a member of the Ash Center’s Allen Lab for Democracy Renovation in addition to working for Harvard University’s Edmond and Lily Safra Center.
Andrei Stupu was a previous Allen Lab for Democracy Renovation fellow at the Ash Center.
The views expressed in this article are those of the author(s) alone and do not necessarily represent the positions of the Ash Center or its affiliates.
Commentary
Allen Lab member Charlie Covit reflects on the After Neoliberalism conference and examines the intersection of artificial intelligence and the future of work, arguing that AI forces a democratic reckoning with the meaning of labor itself and that an economy which generates abundance while stripping citizens of purpose and dignity undermines the very foundation of democratic life.
Media Release
A new book from Harvard scholar Danielle Allen revisits the forgotten British radical movement that helped shape modern democracy.
Media Release
New study published in AI and Ethics introduces a new ethical-moral intelligence framework for AI and finds that leading AI models mimic human moral concern while making decisions that reveal a hidden value hierarchy.
Commentary
Allen Lab member Charlie Covit reflects on the After Neoliberalism conference and examines the intersection of artificial intelligence and the future of work, arguing that AI forces a democratic reckoning with the meaning of labor itself and that an economy which generates abundance while stripping citizens of purpose and dignity undermines the very foundation of democratic life.
Media Release
New study published in AI and Ethics introduces a new ethical-moral intelligence framework for AI and finds that leading AI models mimic human moral concern while making decisions that reveal a hidden value hierarchy.
Article
A new chapter in APSA Preprints by Archon Fung, Winthrop Laflin McCormack Professor of Citizenship and Self-Government and Director of the Ash Center, Bailey Flanigan, former postdoctoral fellow at the Ash Center and co-authors explores how generative AI is reshaping four dimensions of democratic practice—political campaigns, election administration, social movements, and citizen deliberation. The authors argue that AI’s ultimate democratic impact will depend less on the technology itself, and more on how institutions and leaders implement and regulate it.
