The Ethical Need for Watermarks in Machine-Generated Language - CEA - Commissariat à l’énergie atomique et aux énergies alternatives Accéder directement au contenu
Article Dans Une Revue AI and Ethics Année : 2022

The Ethical Need for Watermarks in Machine-Generated Language

Résumé

Watermarks should be introduced in the natural language outputs of AI systems in order to maintain the distinction between human and machine-generated text. The ethical imperative to not blur this distinction arises from the asemantic nature of large language models and from human projections of emotional and cognitive states on machines, possibly leading to manipulation, spreading falsehoods or emotional distress. Enforcing this distinction requires unintrusive, yet easily accessible marks of the machine origin. We propose to implement a code based on equidistant letter sequences. While no such code exists in human-written texts, its appearance in machine-generated ones would prove helpful for ethical reasons.

Dates et versions

cea-03904741 , version 1 (17-12-2022)

Identifiants

Citer

Alexei Grinbaum, Laurynas Adomaitis. The Ethical Need for Watermarks in Machine-Generated Language. AI and Ethics, In press. ⟨cea-03904741⟩
23 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More