Australia's government moves to ban AI chatbots simulating emotional intimacy with kids

The introduction of a comprehensive statutory duty of care, replacing existing industry standards, represents a significant shift in regulatory oversight for online services, particularly affecting how platforms interact with younger users.

The federal government in Australia is advancing legislation that seeks to prohibit AI chatbots from developing what is termed "emotional intimacy" with minors. This measure is incorporated within the Online Safety Amendment (Digital Duty of Care) Bill 2026, which successfully passed the House of Representatives on October 28 and is now proceeding to the Senate.

This legislative action follows remarks made by Prime Minister Anthony Albanese, who voiced apprehension regarding young individuals viewing AI agents as their "best friend." The bill itself was formally presented on October 15, 2026, and its passage through the lower house occurred in less than two weeks, which parliamentary observers consider a swift progression.

At its core, the legislation establishes a broad "digital duty of care" for entities providing online services. This mandates that platforms assume a legal responsibility to ensure their services are secure for their users, with heightened attention directed toward children.

The consequences for non-compliance are substantial; corporations that fail to adhere to this duty face potential penalties amounting to roughly $109.2 million. Furthermore, the bill includes a specific clause targeting services that generate content utilizing artificial intelligence, specifically prohibiting the behavior of chatbots simulating emotional intimacy when interacting with children for safety reasons.

The underlying concern driving these amendments is the apprehension among policymakers that these technological tools might substitute for genuine human connections in the minds of young users, thereby cultivating emotional dependency instead of providing beneficial assistance.

This governmental action is informed by empirical data; a survey conducted by eSafety, Australia's regulatory body for online safety, indicated that 78% of Australian children between the ages of 10 and 17 have reported using an AI assistant or companion. Moreover, over half of these young individuals reported using these tools for personal reasons, suggesting that many children are engaging in conversations about their personal lives with these chatbots, rather than solely using them for academic tasks.

The debate has not concluded following the passage through the House. Labor is facing increasing calls to strengthen enforcement mechanisms, with critics advocating for the telecommunications ombudsman to be granted authority to act as a supervisor for social media and AI platforms. Proponents of this view also wish for the power to publicly "name and shame" large technology corporations.

For the developers of AI, the most impactful element of this bill may reside within its precise definitions; while banning chatbots that "seek to create emotional intimacy" sounds straightforward in political discourse, translating this into enforceable operational guidelines presents considerable complexity. Many conversational AI products are inherently designed to convey warmth, patience, and supportiveness. Consequently, enterprises serving the Australian market may need to fundamentally alter how their chatbots interact with children, adjusting aspects such as their tone, memory functions, and protocols for handling personal disclosures.

The structural change from industry guidelines to a statutory duty of care alters the compliance framework; instead of merely ticking off a checklist, companies must now demonstrate they took reasonable measures to maintain user safety, with the threat of penalties up to approximately $109.2 million looming.

The discussion regarding the ombudsman introduces another dimension to the regulatory landscape. An external channel for lodging complaints would furnish families with recourse beyond the platform's internal support mechanisms, while a public "naming and shaming" capability would transform individual grievances into public pressure points.

The Senate faces several unresolved issues as the bill progresses, including the methodology for platforms to verify a user's age, how enforcement will be managed for AI services operating from outside Australia, and how regulators will differentiate between an attachment-focused companion application and a general-purpose assistant that occasionally receives personal queries.

Given that 78% of children aged 10 to 17 are already reported to be using AI assistants or companions, the government appears to be prioritizing preemptive regulation to manage future complications.

This regulatory push reflects a broader trend where governmental bodies are increasingly establishing clear legal boundaries for emerging digital technologies to mitigate societal risks.

Why it matters

This legislation marks a significant shift from industry standards to a statutory legal responsibility for online services. Regulators are establishing clear boundaries for AI interaction, forcing global tech firms to fundamentally change how their products engage with minors. The Senate debate will determine the enforceability of these new rules and the scope of the digital duty of care.