A federal inquiry has logged more than 420 submissions on algorithmic racism against Indigenous Australians. The Human Rights Commission wants a duty of care to follow.
A federal parliamentary inquiry has received more than 420 submissions from First Nations Australians describing an increasingly toxic online environment, and the Australian Human Rights Commission has used the inquiry to push a specific structural remedy: a statutory digital duty of care that would hold platforms responsible for what their recommender systems amplify.
The submissions, lodged with a parliamentary inquiry into racism, hate and violence directed at Indigenous Australians, describe a daily, persistent experience of racist content being pushed into social media feeds. The volume and consistency of the testimony has, according to reporting this week in the Guardian Australia, moved the policy conversation from platform nuisance toward public health Online racism is significantly impacting mental health, First Nations people say, Guardian Australia.
The harm, as submitters describe it in the material on the public record, is not a string of isolated incidents. It is a steady state. Social media feeds tuned to engagement, the testimony says, treat racist content about Indigenous Australians as a high-performing category and serve more of it. The mental health cost, in the framing used by First Nations submitters, is continuous.
"It's like carrying a bully in your pocket." The line, reported by the Guardian as the way First Nations interviewees put the experience, captures what submitters say distinguishes online racism from its older forms. The abuser does not need to be present, and the abuse does not need the target's participation. The phone does the work.
The Human Rights Commission's response, on the public record, is to argue for a digital duty of care. The construct, reported by the Guardian as a Commission recommendation rather than adopted government policy, targets the design choices that materially amplify racist content: recommendation engines, monetisation rules, and the reward structures that make virality profitable. The framing is structural. It asks government to write a duty platforms cannot discharge with a content moderation team and a transparency report.
The texture of the week has been made visible by a single viral artefact. A self-declared Australian comedian posted a video in which a white woman in a fur coat, her face dotted with white paint, presents herself as "Aunty Lisa" and claims Aboriginal identity after ticking "yes" on an identification form. The clip has, by the Guardian's account, circulated through Aboriginal and Torres Strait Islander feeds for roughly a week and a half. Its function, as several submitters have described it, is illustrative: the kind of content a recommender system surfaces when no one asks it to stop.
The unresolved question is legislative. The Commission has put the duty of care on the table as a recommendation. Government has not, on the present public record, adopted it. The parliamentary committee's report, when it is delivered, will be the first formal test of whether the duty of care moves from recommendation to obligation, and whether the platform response becomes a design question or a press statement.
The 420 submissions on the record are the evidentiary base a duty of care would rest on. The committee's terms of reference, the Commission's recommendation, and the platforms' answers to both are the things now in front of government, and the things to watch.