Australia moves to ban AI chatbots from simulating emotional intimacy

The focus on emotional intimacy

The Australian federal government is moving to restrict how AI chatbots interact with users, specifically targeting those that seek to create "emotional intimacy." Prime Minister Anthony Albanese recently highlighted concerns regarding children who treat these digital agents as their "best friend."

The government argues that these interactions can lead to a dangerous shift where children prioritize non-human relationships over real-world human connections. This concern prompted the inclusion of specific language in the new digital duty of care bill to address these behaviors.

As reported by The Guardian, the bill aims to prevent platforms from hosting material or conduct that simulates intimacy to foster a child's dependence on a chatbot. The goal is to ensure that technology does not replace the fundamental social structures of human life.

New safety standards for platforms

The proposed framework establishes an "overarching obligation" for tech platforms to maintain a safe online environment for all Australians. This replaces several existing online codes and standards with a single, broader duty-of-care requirement. It covers a wide range of services including social media, games, apps, and chatbots.

Under these rules, platforms would be responsible for removing content that promotes specific harms, such as terrorism, misogyny, suicide, and eating disorders. The legislation seeks to move the responsibility of safety directly onto the companies providing the services.

Additionally, the bill requires social media platforms to offer an "opt-out" feature for their algorithms and suggested content. To ensure these features are not hidden, the government accepted amendments requiring them to be prominently available and easy for the average user to find.

Political hurdles and debates

While the bill passed the House of Representatives, its future in the Senate remains uncertain. The Greens support the legislation but are pushing for even stronger protections, including the right for Australians to pursue private legal action against platforms. They are also calling for higher penalties for companies that fail to comply.

The Coalition opposition has stated it will not vote for the bill, accusing the government of creating an "Orwellian" system of censorship. They argue the laws could be used to influence online conversations and limit free speech. Communications Minister Anika Wells has denied these claims, stating the focus is on safety rather than censorship.

Labor is currently attempting to navigate these political divisions to secure enough support to pass the law. The government hopes to have the bill finalized by the end of 2026, though the pathway through the upper house is still being debated.

The role of a tech watchdog

A significant point of contention involves whether to empower the telecommunications ombudsman to act as a watchdog for big tech. Currently, this office does not have the power to help users with issues like hacked accounts or inaccessible Gmail addresses. Supporters argue a watchdog is needed to provide an external complaints mechanism.

Cynthia Gebert, the telecommunications industry ombudsman, suggested that her office could be well-positioned for this role if its remit were expanded. She noted that such a role could be funded by the very industries it regulates. However, the government did not accept amendments to make this expansion mandatory in the lower house.

The Greens are continuing to push for this reform during Senate negotiations. They argue that without a dedicated watchdog, Australians are left with automated responses from big tech companies that often do not resolve actual problems.

What comes next for the bill

The bill is now moving into the Senate, where the Greens have established an inquiry to probe the laws further. This inquiry is expected to report back in late November, which will likely influence the final version of the legislation. The outcome of this inquiry will be a key indicator of how much the laws will be strengthened.

During the debate, the government will need to balance the demands of the Greens for private legal action and a watchdog against the Coalition's concerns over censorship. The final framework will determine how much power the government has to define and police "harmful" content on the internet.

For now, the focus remains on the "emotional intimacy" ban. This specific provision marks a clear intent to regulate the psychological boundaries of AI, setting a precedent for how other countries might handle the social impacts of generative technology.