99% Harmed: Australia Moves Dating Apps Toward Binding Safety Duty

99 percent. In survey research that number almost never appears, and that is why it lands so hard. Of 1,008 heterosexual Australian women who had used dating apps in the prior three months, all but a handful said they had encountered at least one harmful behaviour. Not over years of dating, but in a single season of matching, messaging and meeting ^{1 2}.

The finding comes from Understanding Tech-Based Abuse of Women on Dating Platforms, a Mindaroo-funded report prepared by the Behavioural Insights Team and released around 22 September 2026. Its definition of harm is deliberately wide. It counts overt abuse together with more normalised manipulation, and argues that constant low-level degradation has become so routine that many women blame themselves ^{2 3}. That choice explains both the power of the headline and the argument about what should follow.

Inside the near-total exposure, the breakdown is grimly detailed. About 95 percent reported psychological or emotional harm. About 89 percent experienced monitoring or surveillance. About 88 percent received unsolicited sexual images or videos or were pressured to send sexual content. Almost six in ten reported fear-based controlling behaviour including threats of physical harm ². The spectrum runs from unsolicited graphic messages, photos and videos through negging, love bombing, gaslighting and persistent contact to monitoring, stalking and explicit threats ^{1 2}.

What women do with that experience may matter as much as the numbers. The research finds they largely protect themselves: unmatching quickly, screening profiles for warning signs, withholding identifying information, warning and coaching each other rather than relying on platform tools. Only about 22 percent of those who received threats reported them to platforms. Among those who did report generally, 19 percent felt their complaint was barely or not at all taken seriously ². Design features such as anti-screenshot tools can hinder documentation, leaving threatening behaviour hard to prove and easy to dismiss.

Canberra wants to move that reality from persuasion to legal obligation. The exposure draft of the Online Safety Amendment Digital Duty of Care Bill 2026, released on 8 September 2026 to amend the Online Safety Act 2021, would require persons responsible for online services to ensure, so far as reasonably practicable, a safe online environment for people in Australia ^{12 14}. The draft follows a May 2026 issues paper and the 2024 review of the Act, and consultation by email closed at 12pm on 22 September 2026 ^{12 13}.

In practice the duty is systems-based and proactive. Providers would have to identify foreseeable risks, assess them, implement and document mitigations around design features, algorithms and artificial intelligence, keep those measures effective over time and meet transparency requirements ^{17 20}. The protection covers seriously harmful material and conduct including child sexual exploitation, sexual violence, explicit threats of physical violence and harassment, with extra protections for children covering pornography, hostile attitudes toward women and gender equality, eating-disorder content, glorification of crime or dangerous stunts, and abuse and bullying.

The scope is deliberately broad. It spans nine statutory categories of online service, including relevant electronic services such as dating apps as well as social media, messaging, games, search, hosting, app stores, devices and AI providers, and it would replace the Basic Online Safety Expectations and existing industry codes and standards ^{20 28}. Australians would also gain a right to opt out of algorithmically recommended feeds and see content only from accounts they follow ^{15 29}. A voluntary Online Dating Code signed by major apps already exists. The duty would make risk mitigation legally binding rather than voluntary ^{40 41}.

Enforcement would sit with the eSafety Commissioner with expanded powers, backed by civil penalties of up to 109.2 million Australian dollars and a 12-month transition after commencement ^{19 28}. Communications Minister Anika Wells has framed the package as giving users choice over feeds and as a reckoning for large technology companies, while Treasurer Jim Chalmers has vowed a crackdown on monetised misogyny and committed that dating apps will be covered, with legislation aimed before the end of 2026 ^{15 25}. The government has pointed to introduction in October, though no final bill has been introduced and the draft still carries a blank number ¹⁴.

The politics will decide whether that promise becomes enforceable law. Labor needs Senate support. The Greens welcomed the draft as a step toward holding platforms accountable but said more work is needed, arguing the social-media age ban has not made platforms safer and positioning themselves as the likely pathway ³⁹. That leaves a negotiation over strengthening accountability while answering objections about speech, scope and ministerial power.

What remains unresolved is the text platforms will actually have to obey. The precise obligations for dating apps around prevention, detection, complaint handling, safety-by-design and transparency are not yet fixed. Full methodology behind the headline rate, including the Experience Map and the GADA scale for measuring tech-based abuse, remains only partly public beyond the 1,008-woman heterosexual sample and three-month window ³. Without a negotiated Senate deal, the near-universal harm documented over three months stays inside a voluntary code, under-reported, poorly remedied and managed by women themselves ^{2 38}.