Online hate is often discussed as if it were only a content-moderation problem: remove the worst posts, suspend the worst users, publish a transparency report and move on. But the evidence heard in Australia’s antisemitism and online-hate debates points to something deeper. The issue is no longer only what people say online. It is how digital platforms organise attention, reward outrage and shift social risk onto communities already under pressure.
SBS reported that an Australian inquiry heard warnings about unprecedented levels of online hate speech, with antisemitism a central focus. The Australian context matters, but the problem is global. Democracies are still trying to decide what it means for private platforms to operate spaces that function like public infrastructure while being governed by opaque corporate systems.
Why this is not only about one group
Antisemitic abuse has specific histories and contemporary consequences, and it should not be flattened into a generic civility problem. At the same time, the systems that spread antisemitism often spread racism, misogyny, Islamophobia, anti-Asian abuse, homophobia, conspiracy theories and harassment of journalists, politicians and public servants. The platform machinery is shared even when the targets differ.
That creates a policy dilemma. A society has to recognise the specific harm to Jewish communities while also building rules that protect other groups and democratic participation. If people withdraw from public debate because abuse becomes normal, free expression becomes formally available but practically unequal.
The algorithmic layer
Older hate-speech debates often imagined an individual speaker and an individual listener. Platforms changed the scale. Recommendation systems, engagement metrics, reposting tools and group dynamics can amplify material that would otherwise remain marginal. A hateful post can become a signal around which strangers organise. Harassment can become a campaign. A conspiracy can become an identity.
That does not mean every platform is responsible for every user action. But it does mean the architecture matters. The design of feeds, reporting tools, friction before sharing, enforcement speed and appeal systems can either reduce harm or make it easier to scale.
The free-speech counterargument
The strongest counterargument is real: governments should not use hate-speech concerns as a blank cheque to police political disagreement or unpopular views. Democracies need room for argument, offence, satire, religious debate and criticism of governments and institutions. Overbroad rules can be abused.
But that concern should lead to precise safeguards, not resignation. The question is how to distinguish robust debate from targeted harassment, incitement and dehumanising abuse. It is also how to ensure decisions are transparent enough that users and communities can understand why content stays up or comes down.
Why platform responsibility is hard to enforce
Regulators face a moving target. Platforms change features quickly. Content travels across mainstream networks, encrypted channels, fringe forums and short-video feeds. Local law struggles against global companies. Even when platforms remove content, they may not change the incentives that made the content travel.
This is why inquiries often become frustrated with transparency. Policymakers want to know what companies knew, when they knew it, how quickly they acted, how enforcement differed by language and community, and whether business incentives made harm more likely. Without that information, public debate relies too much on company assurances.
What a serious response would look like
A credible response would combine several layers. First, stronger transparency obligations about moderation systems and risk assessments. Second, better support for targeted communities and individuals, including reporting pathways that do not require victims to repeatedly expose themselves to abuse. Third, platform design changes that reduce the speed and reach of coordinated harassment. Fourth, digital-literacy work that treats conspiracy thinking and dehumanising language as civic risks, not only personal bad behaviour.
None of this is simple. But doing little is also a choice. It leaves the rules of public life to companies whose primary duty is not democratic health.
The wider lesson
Australia’s inquiry matters because it captures a larger democratic transition. The public square has moved into privately owned systems, but the social costs of failure are borne by communities, schools, synagogues, mosques, newsrooms, workplaces and families. If democracies want open debate, they have to protect the conditions that allow people to participate without being targeted out of public life.
The takeaway is that online hate is not just a moderation backlog. It is a governance problem. Platform accountability must be precise enough to protect speech and serious enough to protect people.
Sources: SBS News on the online-hate and antisemitism inquiry, Australia's eSafety Commissioner.
