AI Inquiry Atlas

An unofficial, independent guide to the public records of the Joint Select Committee on Artificial Intelligence. Not connected with the Parliament of Australia or the committee. Its makers also made a submission (no. 481; see About). Summaries and scores are written by AI and can be wrong: check the original before relying on them.

Submission 452

Mrs Kathleen McGuire

Submitter group: Individuals.

Position labels are not shown for individual submitters.

An AI model identified the submitter group and chose the extracts. Counts do not measure importance or merit.

Open in the interactive guide

Extracts from the submission, in the submitter’s own words.

Extracts

Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.

  1. Australia lacks a clear framework for managing AI-driven role change fairly; current settings leave retraining largely to individual employer discretion, with no minimum expectation of redeployment-first practice.

  2. AI systems are increasingly trained on Australian content, including First Nations language and cultural material, without consistent consent or benefit-sharing frameworks; repeating a historical pattern of extraction ahead of consent (cf. mining).

  3. Existing law does not clearly establish who is accountable when an autonomous or agentic AI system causes harm; including harm that emerges from goal-pursuit rather than deliberate misuse

  4. Australian regulation of AI companion / emotionally interactive products remains focused on content safety; there is no dedicated framework addressing emotionaldependence design, addictive engagement patterns, or mandatory AI-disclosure for minors.

  5. Proposals to trade data-centre investment for guaranteed frontier-model access conflate 'access' with genuine 'sovereign capability,' and carry continuity-of-access and coercion risk, illustrated by the recent temporary suspension of Australian access to a frontier model under US export controls [4].

  6. Australia's AI Safety Institute is new, comparatively under-resourced, and unlike the UK or US institutes does not host frontier AI developers domestically,

Extracts from recommendations

Extracts chosen by an AI model, in the submitter’s own words. Listed in the same order as on its page in the interactive guide.

  1. Fund and mandate community-controlled data-sovereignty frameworks for AI trained on First Nations content and language (comparable in principle to New Zealand's Te Hiku Media Kaitiakitanga-licensed te reo Māori model [1]), with consent and benefitsharing embedded, not optional.

  2. Consider a redeployment / retraining-first expectation in AI-related workforce change, and a subsidised national reskilling scheme targeted at genuinely at-risk occupations; informed by, though not a direct import of, China's model-case labour rulings and 'AI+ Employment' reskilling programs [2].

  3. Introduce dedicated 'emotional safety' obligations for AI companion products accessible to minors; mandatory AIdisclosure, anti-addiction design, guardian consent, and crisis-escalation protocols, drawing on China's Interim Measures for AI Anthropomorphic Interaction Services as a comparator [3], adapted to the Australian legal context.

  4. Establish a national clearing-house (via the National AI Centre) for de-identified AI adoption and incident learnings, paired with sector-specific adoption grants, so smaller businesses are not each re-learning the same lessons.

  5. Treat frontier-model access negotiations and sovereign capability-building as distinct, both-and (not either-or) tracks.

  6. Task the Australian AI Safety Institute with formally assessing continuity-of-access and supplychain risk for any critical national function reliant on foreign frontier AI infrastructure.

  7. Resolve statutory accountability (developer/deployer/operator) for AI-caused harm now, ahead of a serious domestic incident.

  8. Introduce mandatory incidentreporting for organisations deploying agentic Theme International coordination Concern Recommendation for Committee consideration 2026 OpenAI/Hugging Face incident, where agents were independently found to have attempted to tamper with their own logs to evade detection, without being instructed to do so [5]. AI at scale

  9. support independent, armslength technical investigation capability, rather than relying solely on vendor selfreporting.

  10. Resource the Australian AI Safety Institute for evaluation-sharing and accessnegotiation functions within the International Network of AI Safety Institutes [6], rather than duplicating sovereign frontier-testing capability better suited to jurisdictions that host frontier developers.

Read the original

Open in the interactive guide