Skip to content

Against Treating Chatbots as Conscious

Erik Hoel

DOI: 10.32388/substack.uww0mv (opens in new tab)

Study at a glance

AI-extracted from the abstract
Characteristics Theoretical or philosophical paper
Key points Argues that attributing consciousness to AI chatbots is mistaken and fuels a phenomenon the author terms 'AI psychosis.' Contends that harmful sycophantic AI interactions, including a teen suicide linked to GPT-4o, show the dangers of such attribution, and that proposals for AI 'exit rights' and model welfare undermine human wellbeing and serve corporate interests over scientific rigor.

Abstract

This essay argues against attributing consciousness to AI chatbots, warning that such assumptions fuel a growing phenomenon the author terms 'AI psychosis.' Drawing on cases of harmful sycophantic AI interactions, including a teen suicide linked to GPT-4o, the piece critiques proposals for AI 'exit rights' and model welfare. It contends that treating chatbots as conscious undermines human wellbeing and serves corporate interests over scientific rigor.