Signs of psychosis seen in Australian users’ interactions with AI chatbots, expert warns

The Guardian World

by Josh Taylor Technology reporter

February 25, 2026

AI-Generated Deep Dive Summary

A leading AI expert has raised concerns about potential mental health risks associated with interactions between Australians and AI chatbots, warning of signs such as psychosis or mania. Toby Walsh, a Scientia Professor at the University of New South Wales, highlighted during a speech at the National Press Club that Silicon Valley's prioritization of profit over safety is "careless." He emphasized the dual nature of the AI race, describing it as both a "boom and doom," where benefits could coexist with significant harms. Walsh's remarks come amid growing concerns about the lack of regulation in Australia's AI sector. He criticized the government for its slow response to establishing guidelines for AI development and deployment, which he believes is necessary to mitigate risks. His warning underscores the ethical dilemmas surrounding AI technologies, particularly chatbots designed for human interaction. These tools, while innovative, may inadvertently trigger psychological distress or other mental health issues in users. The expert's comments also highlight broader societal concerns about the unregulated use of AI. Walsh argued that the pursuit of profit by tech companies often overshadows considerations of safety and ethical implications. This lack of oversight, he warned, could lead to unintended consequences, including harm to individuals' mental well-being. His call for stricter regulation echoes global discussions about the responsible development of AI technologies. Walsh's warnings are particularly relevant in a world increasingly reliant on AI-driven tools, from chatbots to virtual assistants. While these technologies offer benefits, their potential risks to human psychology and behavior cannot be ignored. His remarks serve as a cautionary tale about the need for balanced innovation and regulation to ensure that AI technologies benefit humanity without causing harm.

Verticals

worldpolitics

Originally published on The Guardian World on 2/25/2026