Australia’s House Passes a Safety Duty for Chatbots

Human Verified 0

Share to: , and

A parliament building at dusk

Australia’s House of Representatives has passed a Digital Duty of Care bill that would require online services, including games, apps and AI chatbots, to protect users proactively, with penalties of up to $109.2 million for failure to comply. The bill passed the House on Wednesday 28 October 2026 and now faces a Senate inquiry detour on 9 November before the upper house considers it, so it is not yet law.

What The Duty Would Require

The bill sets minimum safety standards for online services, requiring them to protect Australians from harmful material and conduct, and to ensure their features and tools do not cause harm to users. Platforms would have to protect under-18s from harmful material, including content that glorifies crime or stunts, causes mental distress, encourages eating disorders or promotes hostile attitudes toward women.

Two further measures show how broad the bill is. Social media services would have to let users over 16 choose whether recommended content or only the pages and people they follow form their primary feed, under a proposal called My Feed, My Way. The feed choice is aimed at giving users, not recommendation systems, control over what they see first.

The eSafety Commissioner would also gain power to remove nudify apps and websites, while the required response time for cyberbullying and abuse complaints would fall from 48 hours to 24. Both measures extend the commissioner’s existing powers rather than creating a new regulator.

An amendment from independent MP Kate Chaney on AI chatbots was absorbed into the bill, according to AAP, addressing so-called attachment hacking: chatbots fostering dependence or sycophancy in place of human relationships. The Greens, meanwhile, want stronger penalties, a sign that the Senate debate may push the bill further rather than watering it down.

A First For Chatbots

Chatbots are the most novel part of the duty, because most existing online safety laws were written for social media feeds and video platforms, not for conversational assistants. Putting AI chatbots inside a safety duty, with nine-figure penalties behind it, would make Australia one of the first major jurisdictions to regulate chatbot safety by law rather than by company policy.

That framing matters for companies building chatbots globally. A duty that covers design choices, such as whether a bot flatters a user into dependence, goes beyond banning harmful content after the fact, and it will force product teams to document safety decisions before launch. Other countries watching the Senate debate will be looking for that detail.

The bill’s supporters present it as the next step in Australian online safety reforms, including the social media minimum-age law. Communications Minister Anika Wells has said parents want to know their children are safe when playing games, chatting, learning or creating, and the government argues that safety standards should apply to digital services as for food, cars and children’s toys.

A parent and teenager looking at a phone together

Why It Matters

If the Senate passes the bill, it will change the compliance calculus for every large platform operating in Australia, and for AI companies whose chatbots are used by Australian teenagers. A $109.2 million penalty is large enough to reach board level, and a proactive duty is harder to satisfy with after-the-fact moderation than a reactive takedown rule.

The chatbot child-safety law signed in New York shows the same pressure building in the United States, but Australia’s bill is broader, covering games and apps as well as chatbots. New York’s law focuses on chatbots for children, while Australia’s covers a wider set of services.

The Senate stage is where the bill’s final shape will be decided, and the Greens’ push for stronger penalties suggests the upper house may not simply wave it through. For now, the House vote is a clear signal of direction: Australia intends to regulate AI chatbots as safety-critical services, not as neutral tools.

The argument over how risky AI models should be governed will now play out, in part, in parliament rather than only in company policy documents. Its final form will depend on negotiations in the Senate, where the Greens are pushing for tougher penalties.

Background

The Albanese government introduced the Digital Duty of Care bill to Parliament on 15 October 2026, and the House passed it on 28 October, with a Senate inquiry detour scheduled for 9 November before the upper house votes.

Source: Minister’s release, AAP report

Share to: , and

Faheem R. Khattak
Founder & Senior Editor

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

✓ 10+ Years Covering Google Search & AI

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.