Portfolio — 21 July 2026

The Attorney-General released a package of AI safety priorities framed as a whole-of-government agenda, with the centrepiece being a proposed Digital Duty of Care that would require AI companies to embed safety by design and proactively address potential harms before they materialise [TA-260720-attorn-736d18c12ecd]. The announcement is explicitly cross-portfolio in character: the Minister for Employment and Workplace Relations nominated AI safety in the workplace as a priority for the tripartite Artificial Intelligence Workplace and Employment Forum, and the Assistant Minister for Productivity, Competition, Charities and Treasury will examine consumer-law options to address emerging risks including retail surveillance pricing and agentic commerce [TA-260720-attorn-736d18c12ecd].

Within the Attorney-General's own portfolio, two further commitments stand out. The Attorney-General announced that the government will consult on a second tranche of privacy reform to modernise and simplify Australia's personal data protection laws — a signal that the first tranche of reform has now been legislatively settled and that attention shifts to a broader rationalisation of the privacy framework.

The Attorney-General also committed to developing a framework to regulate automated decision-making within federal agencies, targeting fair, accurate, and transparent outcomes in Commonwealth service delivery.

On harmful content, the PM media release noted that existing criminal offences already prohibit sharing non-consensual deep-fake sexual material, and confirmed the government is actively working to ban nudify apps — connecting the AI safety agenda to the government's existing online-safety legislative posture.

This release extends a pattern running across the portfolio in the days prior [TA-260720-attorn-736d18c12ecd]. On 16 July, the Attorney-General closed a text-and-data-mining copyright exemption and introduced a modern-slavery offence — moves that established the portfolio's disposition toward constraining AI-adjacent commercial freedoms in favour of rights protections.

Mandatory data-centre standards announced in the same period add an infrastructure layer beneath the safety agenda. Today's announcement builds the regulatory superstructure: duties on industry, protections for consumers and workers, and government-agency accountability — all anchored to the AI Safety Institute and associated research partnerships as the technical foundation [TA-260720-attorn-736d18c12ecd].

The cumulative picture is a portfolio deliberately assembling a multi-layer AI governance architecture — moving from infrastructure standards and copyright constraints to industry duties, consumer protections, workplace safeguards, and government-agency accountability within a compressed legislative and policy window.