[论文解读] Generative Ghosts: Anticipating Benefits and Risks of AI Afterlives

Meredith Ringel Morris, Jed R. Brubaker|arXiv (Cornell University)|Jan 14, 2024

Law, AI, and Intellectual Property被引用 12

一句话总结

本文定义 generative ghosts，绘制其设计空间，讨论潜在收益与风险，并提出在安全、互利的 AI afterlives 之后生活方面的主动 AI/HCI 研究议程。

ABSTRACT

As AI systems quickly improve in both breadth and depth of performance, they lend themselves to creating increasingly powerful and realistic agents, including the possibility of agents modeled on specific people. We anticipate that within our lifetimes it may become common practice for people to create custom AI agents to interact with loved ones and/or the broader world after death; indeed, the past year has seen a boom in startups purporting to offer such services. We call these generative ghosts, since such agents will be capable of generating novel content rather than merely parroting content produced by their creator while living. In this paper, we reflect on the history of technologies for AI afterlives, including current early attempts by individual enthusiasts and startup companies to create generative ghosts. We then introduce a novel design space detailing potential implementations of generative ghosts, and use this analytic framework to ground discussion of the practical and ethical implications of various approaches to designing generative ghosts, including potential positive and negative impacts on individuals and society. Based on these considerations, we lay out a research agenda for the AI and HCI research communities to better understand the risk/benefit landscape of this novel technology so as to ultimately empower people who wish to create and interact with AI afterlives to do so in a beneficial manner.

研究动机与目标

• Define the concept of generative ghosts and distinguish them from griefbots and other post-mortem technologies.
• Identify key design dimensions that shape generative ghosts and their societal impact.
• Assess potential benefits for representees, the bereaved, and society.
• Enumerate ethical, legal, and safety risks and outline mitigations.
• Propose a cross-disciplinary research agenda to enable safe and beneficial AI afterlives.

提出的方法

• Develop a design-space framework with dimensions: provenance, deployment timeline, anthropomorphism paradigm, multiplicity, cutoff date, embodiment, and representee type.
• Differentiate first-party vs. third-party ghosts and pre-mortem vs. post-mortem deployments.
• Clarify the concept of generative ghosts as a specialized form of generative clones.
• Contrast reincarnation vs. representation metaphors to analyze user perceptions.
• Review related work across post-mortem technology and AI afterlives to ground the framework.
• Propose a research agenda integrating AI safety, HCI, end-of-life planning, law, and religion.

实验结果

研究问题

• RQ1 What constitutes a generative ghost and how does it differ from existing griefbots and digital legacies?
• RQ2 What design dimensions determine the capabilities, perceptions, and societal impacts of generative ghosts?
• RQ3 What are the potential benefits and risks for individuals, the bereaved, and society?
• RQ4 What research directions and safeguards should guide the development and deployment of AI afterlives?

主要发现

• Generative ghosts extend beyond static memorials by capable content generation and potential agentic interactions.
• A design-space with seven dimensions (provenance, deployment timeline, anthropomorphism, multiplicity, cutoff date, embodiment, representee type) shapes outcomes.
• Ghosts may be first-party (planned by the representee) or third-party (created by others), with or without consent.
• Potential benefits include comfort for the bereaved, continuity of personal and cultural heritage, and novel educational or cultural applications.
• Risks span mental health (attachment, accommodation, second losses), reputational/privacy concerns, and broader socio-cultural effects.
• The paper outlines a multi-disciplinary research agenda to maximize benefits while mitigating risks.