Value Engineering for Autonomous Agents – Position Paper
[EN]Ethics in Artificial Intelligence is a wide-ranging field which encompasses many open questions regarding the moral, legal and technical issues that arise with the use and design of ethically-compliant autonomous agents. Under this umbrella, the computational ethics area is concerned with the fo...
| Autores: | , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión enviada para evaluación y publicación |
| Fecha de publicación: | 2021 |
| País: | España |
| Institución: | Consejo Superior de Investigaciones Científicas (CSIC) |
| Repositorio: | DIGITAL.CSIC. Repositorio Institucional del CSIC |
| OAI Identifier: | oai:digital.csic.es:10261/235831 |
| Acceso en línea: | http://hdl.handle.net/10261/235831 |
| Access Level: | acceso abierto |
| Palabra clave: | Computational ethics Value engineering Value alignment Normative systems Multiagent systems Philosophical foundations of values in AI |
| Sumario: | [EN]Ethics in Artificial Intelligence is a wide-ranging field which encompasses many open questions regarding the moral, legal and technical issues that arise with the use and design of ethically-compliant autonomous agents. Under this umbrella, the computational ethics area is concerned with the formulation and codification of ethical principles into software components. In this position paper, we take a look at a particular problem in computational ethics: value engineering in autonomous agents. This work aims at building the philosophical foundations that a future model of value engineering should be based on. The main points of our proposal are: (1) values are introduced into agents as goals that ground the meaning of those values; (2) norms are the means to steer an agent society towards beneficial outcomes, and hence should be used to promote values; and (3) autonomous agents can negotiate over norms to align them with their values, in an exercise of value aggregation. Finally, we argue that our proposal does not endow software agents with moral agency, as there is always a human team responsible for deciding which values should be encoded and the meaning they take, i.e. the form of the grounding goals. We believe that this position paper accounts for a solid philosophical foundation for a future formal model of values in autonomous agents, and provides the starting points for work in that direction. |
|---|