Australia is world-leading in taking active measures to keep people safe online – home to the world’s first dedicated online safety regulator, the eSafety Commissioner, and the first country to introduce enforceable industry codes requiring platforms to tackle harmful content at scale.
And now, a newly released federal government issues paper proposes a “digital duty of care”, which would require social media platforms to take reasonable steps to prevent foreseeable online harm.
The proposal signals Australia’s position that it is platforms, not just individuals, who should be responsible for actively preventing online harms.
At the heart of the proposed digital duty of care is the principle that social media platforms should be “safe by design”.
But what does that mean in practice – especially for those who are most at risk? Our research with women and gender-diverse Australians offers six concrete recommendations for what safety by design could look like in practice.
Who bears the brunt of online abuse?
One in two Australian adults have experienced online abuse in their lifetime. Women and gender-diverse people are disproportionately targeted, experiencing harassment, non-consensual image sharing, impersonation, stalking and identity-based abuse at far higher rates than others.
Yet these groups are rarely involved in envisioning what safer platforms could look like. So, we asked them: what would safer social media look like to you?
We worked with 75 Australian women and gender-diverse social media users, and 21 experts in platform safety, digital policy and content moderation, to understand how existing safety features are falling short.
Here’s what they told us – and how it compares with the current Australian proposal for a digital duty of care.
1. Make abuse reports actually work. Abuse rarely fits a single category – without context, platforms don’t handle the reports well. A message that reads as innocuous to a stranger may be a clear threat to someone who knows their abuser. But without that context, platforms have no way of knowing.
Users want clearer processes that capture the full picture, smarter triage that prioritises urgent cases, and timely updates on what happened to their report. This fits well with what the digital duty of care proposes: platforms should have accessible complaint mechanisms and respond within 24 hours for serious issues.
2. Harmful content should be harder to share in the first place. Once someone shares intimate or sensitive content without your consent, it quickly spirals out of control. Australia’s proposal suggests platforms should prevent the upload of seriously harmful content such as image-based abuse, or detect and remove it.
Users in our research said they want prompts that encourage people to pause before sharing, technical measures that prevent screenshots or downloads, and real-time alerts showing when and where their content is being accessed.
3. Make bans harder to evade. If you block a user, they can create new accounts in minutes, facing few real barriers. The digital duty of care flags that anonymous account systems may need redesigning to prevent foreseeable harm.
As we found, users want layered verification – such as requiring a unique phone number or introducing delays before new accounts become active – that adds friction to repeat account creation, but not mandatory ID checks for everyone. This would protect those without formal ID, those escaping unsafe homes, or those who rely on anonymity to stay safe.
4. Harmful content should be caught before it spreads. Automated systems routinely miss culturally specific abuse and coded language. Content should be detectable before it is shared, and easy for bystanders – not just victims – to flag.
The users in our research recommended pairing automated detection with human moderators trained in cultural nuances, which is precisely the kind of effective content moderation system the proposed duty of care requires.
5. Recognise campaigns, not just individual posts. Abuse is often a sustained campaign, even when each message seems minor alone. The duty of care proposal requires platforms to mitigate reasonably foreseeable harms – which means looking beyond individual incidents.
Platforms should connect reports over time, identify patterns, and act before harm escalates, with independent audits to ensure these systems are never weaponised against the people they are meant to protect.
6. Surface safety tools before harm happens. Most users discover safety features only after something has gone wrong. Australia’s proposal envisions “empowering” users – but empowerment means more than adding features. It means the platform should offer the right tool at the right moment, rather than bury it in a settings menu that only the most determined users will ever find.
The real test
The proposed digital duty of care is a significant step in the right direction. But “safe by design” will only deliver if it works for everyone. As our research shows, those most affected already have clear, practical ideas about what would make platforms safer.
The opportunity now is to design with them – so safety is built in from the start.
Until the proposed digital duty of care is rolled out, it is up to all of us to look after each other. We can report harmful content, pause before we post and ask: is it true? Is it kind? Is it fair? And we can be active bystanders – commenting when we see something harmful, or offering support to those experiencing abuse.
We all have a role to play. From governments, to platforms, to everyday people – it is up to all of us to create a safe digital society, one that we can all be a part of.
