Before ParliamentTech & Online Introduced in the House of Representatives, 15 Oct 2026

Tech firms would have a legal duty to keep users safe

Apps, social media, search engines and AI tools would have a new legal duty to stop serious harm before it happens. The online safety watchdog, called eSafety, would get power to have fake nude apps removed. Fines could reach almost $22 million.

Official name: Online Safety Amendment (Digital Duty of Care) Bill 2026

Who it affects

• Social media, apps, search engines and AI services: They would have a legal duty to keep users safe. They would need yearly risk checks and could face fines of up to $21.84 million.
• Children and teenagers: Services would have to protect them from more kinds of harmful posts. Social media would have to keep kids under 16 away from some features, such as feeds that pick posts for you.
• People hit by bullying, abuse or fake nude images: The watchdog could act faster. In some cases you would not need to report it to the app first.
• Social media users over 16: The minister could make apps give you tools to choose what shows up in your feed. The minister says this is the plan.
• Uni researchers: Researchers approved by the watchdog could get safety data from tech companies and use fake accounts to test apps.

Why it exists

The Online Safety Act 2021 started in January 2022. It set up the eSafety Commissioner. That is the official whose job is to keep Australians safe online. The office is called eSafety for short.

Under that law, harm is mostly dealt with after it happens. People have to report bad posts. Or they have to take steps to keep themselves safe.

The government says serious harm is still common. eSafety found 7 in 10 kids aged 10 to 15 had seen harmful posts online. These included hate aimed at women, risky dares, fight videos and posts that push eating problems or suicide.

More than half of the kids asked said they had been bullied online. A study led by the University of Adelaide looked at people under 18. It found 1 in 25 had been sexually abused using AI, or knew someone who had. One example is AI being used to make fake nude pictures of a real person.

The law said there had to be a review of how well it was working. Delia Rickard did that review. She gave it to the government in October 2024. She said the law was no longer strong enough.

She made 67 ideas for change. One was a single big duty of care for online services. In April 2026 the government replied. It said it would deal with 64 of the 67 ideas in some way. Some it would act on fully, some only in part. Others it would look at more closely before deciding.

Anika Wells is the Communications Minister. That is the member of the government in charge of phones and the internet. She brought in this bill. She said it puts the job of stopping harm on the people who design the product.

Fake nude apps removed

Part 1 of the bill gives eSafety a new power over nudify apps and websites. These let people make real-looking fake nude pictures of someone from a normal photo or video.

eSafety could tell an app store or search engine to remove such an app or site. It could also tell them to take down links and ads for it.

eSafety must have good reason to believe the app is mainly for making fake nudes. That counts if it was built for that, or if that is mostly what people use it for.

The app store or search engine would have 24 hours to act, unless eSafety gives more time. Ignoring the order could bring a big fine. For a company, that could be up to about $10.9 million. The company can ask a review body to look at the order again. That body checks choices the government makes.

Faster takedowns of bullying

Here is how it works now. A child might be bullied online. Or an adult might be abused online in a way meant to cause serious harm. Either way, the victim must first complain to the app. eSafety then waits 48 hours before it can order the post taken down.

The bill cuts that wait to 24 hours.

eSafety could also skip the first step in some cases. One is when the app has no working way to complain. Another is when complaining to the app could put the person in more danger.

It could also skip that step when someone puts a post back up. That means eSafety ordered the same post down in the last 6 months, and it has appeared again. This counts even if the post was cropped or changed.

Say a group of students keeps reposting a cruel edited photo of a classmate. Once eSafety has ordered it down, it could order each repost down straight away.

Links pulled from search

eSafety can already tell search engines to stop linking to the worst illegal posts. The bill would let it do the same for bullying and abuse. It could also do it for private sexual images shared without the person's OK.

eSafety could do this as soon as it orders the post taken down. That makes the post harder to find while it is being removed. Search engines would have 24 hours.

Once the post is gone, there is nothing left to link to. So eSafety must then cancel the link order. It must also cancel it if it is no longer needed for another reason.

Bigger fines and sharing information

Most fines for ignoring eSafety's orders would go up about twelve times, from $182,000 to $2.184 million. That is the top fine for a person. A court can fine a company up to five times as much, which is about $10.9 million.

The tech industry has its own content rules. Some were written by the industry and some were set by eSafety. Fines for breaking them would double, from $10.92 million to $21.84 million. Those rules would stay for one more year, until the new duty starts.

Some of these fines can also hit ordinary people. Say someone posts abuse aimed at an adult, and eSafety orders them to take it down. If they refuse, they could face the bigger fine. The same goes for someone who posts a private sexual image of another person without their OK.

The Parliamentary Library says a top fine of $2.184 million may seem too big for one person. The Library works for all members of Parliament and takes no side.

Today eSafety can share information with only a short list of national government bodies, such as the Australian Federal Police. The bill would let it share with any national government body, if that helps the body do its job.

The duty of care

Part 2 is the biggest change. Anyone who runs an online service would have to take the right steps to make it safe. This is the digital duty of care.

It covers social media, search engines and app stores. It covers companies that store websites, internet providers and AI tools that make text or pictures. The minister could say some low-risk services do not have to follow the duty.

The bill lists the most harmful posts and acts. Services must keep everyone in Australia safe from them. Most of the things on the list are already crimes or banned in Australia. The bill adds posts that urge or teach violence against people because of their gender.

Kids need more protection than adults. So services must also keep kids away from things that are harmful to them. These include porn and posts that push eating problems or hate toward women. They include posts that praise crime or risky stunts. They also include abuse, bullying and anything else that may badly hurt a child.

To do this, a service has to work out if kids are likely to use it. If so, it must take steps to protect them.

A separate law already sets 16 as the minimum age for social media accounts. But kids may still use some apps without logging in. This bill adds an extra rule for social media. Kids under 16 must not get design features that can push people to use the app in harmful ways.

The bill names two kinds. One is a feed that picks posts for you, based on what the app knows about you. The other kind covers features that apps keep for people with accounts. Examples are a feed that never ends, or a count of how many people have seen or liked your posts. The minister could add more features to the list.

Some things are left out of the duty. These include news from real news companies, research, health advice, court cases, and shows such as TV and films. Art and learning are left out too, unless they go against what is good for the public.

The duty does not cover private chats between adults who have agreed to chat, as long as the chats are legal. The government's notes on the bill say big group chats with hundreds of people are not private. So the duty can cover those.

A company cannot get out of a safety step by saying it costs too much. The cost to the company does not count when deciding what steps are right. The top fine for breaking the duty is $21.84 million.

Risk checks, feeds and age checks

Services would have to write a risk check at least once a year, and before any big change. It must list what could go wrong, who could be hurt and how likely it is. It must say what the service will do about it, and how well that works.

The minister could make some apps give users tools to control how features work for them. The minister would choose which apps. One example is choosing what kind of posts you are shown. The government calls this My Feed, My Way. The minister says it would give people over 16 a choice about what they see in their feed.

The government has already put out a draft of this rule for people to comment on. It would cover social media apps that pick posts for your feed. They would have to let you choose your own feed.

Some services may let kids see porn. Those services would have to offer users a choice of at least two ways to check their age, if that is possible. Anyone not checked as an adult must be blocked from the porn.

Some services named by eSafety would need a fair way to complain. It must be open to everyone in Australia, even people who do not use the service. Say someone posts a fake picture of you on an app you have never used. You could still complain to that app.

eSafety could give formal warnings, and orders to fix problems. It could post its warnings on its website. Breaking these rules could bring fines of up to $21.84 million.

Reports, probes and research

Right now there is a list of things apps are asked to do to keep people safe. It is called the Basic Online Safety Expectations. A court cannot make apps follow it. The bill scraps the list after a year.

eSafety could then make apps write public reports on how safe they are. That would let users see how each app is doing.

eSafety would get stronger powers to look into problems. If it thinks someone has facts that would help, it could make them hand over papers. It could make them answer questions under oath, which is a promise to tell the truth.

A person who refuses could get up to 12 months in jail, or a fine of up to $364,000. Journalists would not have to tell eSafety the name of a secret source.

Researchers who work at Australian unis could ask eSafety to approve them. Once approved, they could be given company data for online safety research. This could include how the company's computer systems choose what posts to show. The data must be about online safety, and researchers cannot use it to make money.

Approved researchers and eSafety could also use fake accounts to test apps. These are known as sock puppets. Apps have rules that users must agree to. Researchers could not be sued for breaking those rules if they are honestly doing safety research. They could only chat with real users as much as needed to keep the account open.

Old industry rules end

Part 3 scraps the industry's content rules, which are called codes and standards. The new duty would take their place after a year.

It also updates other laws to match. Some companies build filters that find and block child abuse images. Dealing with those images is normally a crime. Criminal law makes an exception for companies doing this filter work. Right now that exception is linked to the industry rules. The bill links it to the new duty, so the exception carries on.

The privacy law limits how information about people can be shared and used. Police and some other bodies whose job is to make people follow the law are allowed to do more with it. The bill would let eSafety do the same. That way the privacy law would not get in the way when eSafety looks into problems.

Waiting in the House

Part 1 has the new powers, such as removing fake nude apps and faster takedowns. These would start the day after the bill is signed into law. The Governor-General, the King's representative, signs bills into law.

The digital duty of care, and the end of the old industry rules, would start a year after that.

Parliament has two houses, the House of Representatives and the Senate. A bill must pass both to become law.

The bill was brought into the House of Representatives on 15 October 2026. The minister asked the House to start talking about it that day.

Senators are the members of the Senate. The Parliamentary Library says a group of senators is looking at the bill. Anyone could send them written comments by 30 October 2026. The group's report is due by 20 November 2026.

How they voted

Its journey through Parliament

Official sources

• Bill page on aph.gov.au
• Text of the bill
• Explanatory memorandum
• Bills Digest

Open the full interactive page