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 254

Global Institute for Women's Leadership, ANU; Wiyi Yani U Thangani Institute; Victorian Women’s Trust; Professor Ramona Vijeyarasa; and Dr Leonora Risse

Submitter group: Research institutes.

An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.

Open in the interactive guide

Summary

Read the original submission for the full text.

Researchers and advocacy organisations argue AI can advance gender equality but may deepen inequalities without intervention. They recommend gender-responsive workforce transitions, binding risk-based legislation, substantive equality, First Nations data sovereignty, deepfake protections and community safeguards for data centres, while supporting AI adoption. Leonora Risse contributes independently; the submission’s views cannot be attributed to affiliated institutions.

Key points

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. Says AI presents significant opportunities to advance gender equality, but argues that without gender-aware, deliberate intervention AI systems will reproduce and deepen existing inequalities.

    AI presents significant opportunities to advance gender equality alongside Australia’s broader economic, social and technological objectives, including through enhancing economic productivity and efficiency. However, creating a gender equitable AI system will require gender- aware, deliberate intervention. Otherwise, AI systems that are modelled on existing systems will reproduce and deepen existing inequalities.

  2. Says emerging Australian evidence shows that exposure to AI-driven automation is highest in female-concentrated clerical and administrative occupations.

    Emerging Australian evidence shows that exposure to AI-driven automation is highest in female-concentrated occupations, namely, clerical and administrative occupations. Women comprise around 70% of this sector’s workforce in Australia (ABS, 2026)

  3. Says gender-patterned workforce impacts do not mean Australia should avoid adopting AI, but argues for a gender-responsive approach to workforce transition, skills and industry policies.

    These workforce impacts do not mean that Australia should shy away from adopting AI in the economy, but it does mean that we should take a gender-responsive approach to workforce transition, skills and industry policies.

    A program corrected obvious text-recognition errors in this quote for display.

  4. States that Australia has no binding AI-specific federal legislation and relies on voluntary instruments.

    Australia has no binding, AI-specific legislation at the federal level. Its approach rests on voluntary instruments, including the 2019 AI Ethics Principles, the 2024 Voluntary AI Safety Standards and the 2025 Policy for the Responsible Use of AI in Government

  5. Argues that a data and governance gap, particularly for First Nations data created without consent, can leave AI systems reliant on inaccurate data and reproducing harmful narratives.

    This is particularly evident regarding data about First Nations peoples, which has historically been created without their consent or ownership, meaning communities have rarely benefited from its development or application. This data and governance gap can leave AI systems reliant on inaccurate data, increasing the risk they reproduce harmful colonial and sexist narratives, exposing the communities to further harm.

  6. Argues that AI intensifies technology-facilitated gender-based violence, citing research that 98% of deepfake videos are non-consensual pornography and 99% of that depicts women.

    Technology-facilitated gender-based violence (GBV) is being intensified by AI, which can automate harassment, enable stalking and create new forms of perpetrator abuse. Non-consensual deepfake pornography is a particularly acute concern, with research finding that 98% of deepfake videos contain non-consensual pornographic content, and 99% of this content depicts women (eSafety Commissioner, 2024).

Recommendations

Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.

  1. Connect gender and AI policy portfolios by mainstreaming gender considerations across AI policy and integrating AI impacts into gender equality policy, supported by gender impact assessments, gender expertise in governance bodies and cross-government coordination.

    Connect gender and AI policy portfolios by mainstreaming gender considerations across AI policy while integrating AI impacts into gender equality policy, supported by gender impact assessments, gender expertise in governance bodies and stronger cross-government coordination.

  2. Leverage existing gender regulation in AI innovation by requiring gender-responsive risk assessments, targeted funding, regulatory sandboxes, transparency requirements and independent auditing.

    Leverage existing gender regulation in AI innovation by requiring gender-responsive risk assessments, targeted funding, regulatory sandboxes, transparency requirements and independent auditing to shape safer and more equitable AI development.

  3. Use AI proactively to reduce discrimination and support gender equality by investing in such applications, with targeted funding and upskilling for women-led and gender-responsive innovation and entrepreneurship.

    Use AI proactively to reduce discrimination and eliminate inequality and bias by moving beyond harm mitigation and investing in AI applications that can reduce discrimination and support gender equality outcomes, including in gendered-based violence, safety and financial services. This will require targeted funding and upskilling for women-led and gender-responsive innovation and entrepreneurship.

  4. Develop enforceable and complementary governance frameworks across global, national and local approaches, with clear standards, adequate resourcing, monitoring and accountability.

    Develop enforceable and complementary governance frameworks across global, national and local approaches, supported by clear standards, adequate resourcing, monitoring and accountability.

  5. Pursue systemic gender equity across organisations and structures rather than placing the primary onus on individual women to adapt through upskilling or self-protection.

    Pursue equity at a system-level not an individual-level. In other words, “fix systems, not women”, by addressing gender inequality proactively across organisational and structural levels, instead of placing the primary onus on individual women to adapt and respond through upskilling or self-protection.

  6. Support global feminist AI champions by building stronger pathways for feminist and gender equality experts to input into AI development, regulation and consultation, such as fellowships, secondments, advisory bodies and networks.

    Support global feminist AI champions by building stronger pathways for feminist and gender equality experts to provide their input towards AI development, regulation and public consultation. This can include support for fellowships, secondments, advisory bodies and cross-sector networks.

  7. Invest in interoperable gender-responsive standards and tools for AI, including shared principles, regulatory templates, impact assessment tools and complaints mechanisms adaptable across jurisdictions with minimum standards.

    Invest in interoperable gender-responsive standards and tools for AI by developing shared principles, regulatory templates, impact assessment tools and complaints mechanisms that can be adapted across jurisdictions while maintaining minimum standards.

  8. Invest in personnel, governed datasets and sector expertise for bespoke domestic/family/sexual-violence and workplace-harassment reporting/support AI, retaining clear pathways to specialist human support.

    Build specialist data capability to support the development of bespoke AI systems for sensitive reporting and support services by investing in skilled personnel, responsibly governed datasets and sector expertise capable of developing purpose-built AI for contexts such as domestic, family and sexual violence, and workplace sexual harassment. These systems could support reporting, triage and referral while maintaining clear pathways to specialist human support.

  9. Require continuous gender and intersectional impact assessments across the AI lifecycle, from design and training through deployment and ongoing use.

    Require continuous gender and intersectional impact assessments across the AI lifecycle to identify and address potential gendered and intersectional harms from design and training through deployment and ongoing use.

  10. Embed equal and inclusive representation of women in AI development and governance, supported by continuous gender impact assessment and independent accountability mechanisms, education and skills development.

    Embed equal and inclusive representation of women in AI development and governance by ensuring women in all their diversity have equal and meaningful representation in and input towards the development, governance and regulation of AI, supported by continuous gender impact assessment and independent accountability mechanisms, and driving education and skills development.

  11. Consider how AI productivity gains reach workers as higher wages and better conditions, including possible gender differences in recognition and bargaining power.

    The key question, that we highlight for the Committee, is to consider how will these productivity-boosting impacts of AI flow through to workers in the form of higher wages and better working conditions? Owing to occupational or industry patterns and/or differences in bargaining power, there may be gender differences in the extent to which these labour productivity improvements are recognised and acted on.

  12. Consider distribution of AI productivity benefits among workers, capital owners/investors and employers, including the gender composition of those receiving returns.

    This raises another economic consideration for the Committee: How will the productivity benefits of AI be distributed between workers in the form of higher wages, the owners of capital and investors of the AI technologies in the form of returns, and employers/businesses in the form of profit? This distribution question also has an evident gender component, given that most investors and owners of AI technologies, and the businesses adopting the technologies, are men.

  13. Assess firms’ retained AI gains versus wider economic spillovers and whether benefit distribution matches costs, including scarce data-centre resources, training IP and environmental pollution.

    the Committee must consider: To what extent will these productivity benefits be retained internally by firms and to what extent will these gains spill over as positive externalities to the wider economy? Will this distribution of benefits be commensurate to the distribution of costs – where the costs of AI include the competitive usage of scarce resources to run the AI data centres, the utilisation of creative and intellectual property to train AI systems, and potential industrial pollution imposed on the environment?

  14. Examine whether the profitable returns of AI technologies are flowing overseas rather than being reinvested in Australia's innovation, technological and economic ecosystem.

    The Committee may also see value in examining whether the profitable returns of these AI technologies are flowing to overseas countries, rather than being reinvested back into Australia’s innovation, technological and economic ecosystem.

  15. Consider whether Australian talent with the capacity to develop AI products is sufficiently incentivised to stay and invest in Australia.

    It is also worth considering whether the Australian minds with the scientific, technological, creative and entrepreneurial capacity to develop these AI products and infrastructure are being sufficiently incentivised to stay and invest in Australia, rather than be attracted to overseas opportunities.

  16. Design skills accreditation and hiring systems to recognise the transferable skills of workers in highly exposed clerical and administrative occupations, and map those skills to new employment and business opportunities.

    Skills accreditation and hiring systems should be designed to identify and recognise the value of the transferable skills, competencies and professional experience held by workers in highly exposed clerical and administrative occupations. These skills should be mapped against into new employment and business opportunities, including new opportunities being created by AI.

  17. Have education, training and workforce development systems prioritise skills and roles that are less readily automated, across the whole education and training system and workforce.

    Education, training and workforce development systems should prioritise skills and roles that are less readily automated. This approach applies across the education and training system, and the workforce, not merely where jobs are being displaced.

  18. Continue action to gender-balance executive decision-making so AI displacement does not compound unequal decision-making power.

    Ongoing action to make these executive decision-making roles gender-balanced is important for ensuring the job displacement effects of AI do not inadvertently compound gender inequities in decision-making power.

  19. Ensure that new and emerging job opportunities arising from AI do not perpetuate gender bias.

    Ensure that new and emerging job opportunities arising out of AI do not perpetuate gender bias.

  20. Improve workforce culture/practices and education/training pathways in male-dominated AI fields, with strong accountability and ongoing evaluation of gender-equity initiatives.

    Initiatives to keep improving the workforce practices and culture of these sectors, and education and training pathways in traditionally male-dominated fields, is critical. This must include establishing strong accountability mechanisms and ongoing evaluation of initiatives intended to promote gender equity.

  21. Increase women’s input and representation in senior sector decisions and women-owned businesses/ventures to reshape culture, practices and inclusive product/service design.

    Beyond simply focusing on encouraging more women to join these sectors, more focus needs to be placed on improving women’s input and representation in senior decision-making positions within these sectors, and lifting the number of women running their own businesses and ventures in the sector, in order to truly re-shape industry culture, workplace practices, and make product and service design more gender equitable and inclusive.

  22. Analyse AI’s labour-productivity impacts for equitable benefit distribution across sectors, particularly non-market and female-concentrated care, community services and education where measurement is difficult.

    Ensure the productivity benefits of AI adoption are equitably distributed by closely analysing the impacts of AI on labour productivity and recognising the ways that AI investment is improving (augmenting) labour productivity. This is important across all sectors, but particular focus should be placed on the non-market and female-concentrated industries – such as care and community services, and education – where measurements of productivity are already difficult and unreliable.

  23. Strongly support ABS efforts to improve methods for measuring non-market sector productivity, highlighting gender equity and the need to recognise AI impacts.

    Current efforts by the ABS to improve methodologies for measuring productivity in the non-market sector should be strongly supported, highlighting both the gender equity elements as well as the need to recognise the impacts of AI.

  24. Develop First Nations data regulatory frameworks embedding Indigenous data sovereignty principles, ICIP rights and free, prior and informed consent in AI development.

    Develop First Nations data regulatory frameworks that recognise and embed Indigenous data sovereignty principles, Indigenous Cultural and Intellectual Property rights, and Free, prior and informed consent into the development of AI systems.

  25. Invest in First Nations and women-led data platforms that can contribute to AI systems without reproducing existing bias and harm.

    Invest in the development of First Nations and women led data platforms that can contribute to AI systems in ways that do not reproduce existing bias, harm and vulnerabilities that existing data models are contributing to.

  26. Engage First Nations and women’s organisations, including GIWL and WYUT, to convene community participation and help government seek diverse women’s voices in decision-making.

    Engage with First Nations and women’s organisations to auspice community engagement: Existing First Nations and women’s organisations, including the GIWL and WYUT, have the social licence, tools and processes in place to enable community engagement processes that have a convening power to help government seek diverse women’s voices and participation in decision-making.

  27. Enact dedicated, binding AI legislation establishing a risk-based framework, with explicit prohibitions on direct and indirect algorithmic discrimination on grounds including gender, race, disability and age, drawing on the EU AI Act.

    Support the enactment of dedicated, binding AI legislation that establishes a risk-based regulatory framework for AI systems deployed in Australia and includes explicit prohibitions on direct and indirect algorithmic discrimination on the grounds of gender, race, disability, age, and other intersecting characteristics, drawing on the EU AI Act.

  28. Mandate Canada-modelled Gender-Based Analysis Plus, at least for public-sector AI, attending to intersectional harms affecting women of colour, socioeconomically disadvantaged/low-income/precariously employed/migrant/Indigenous/disabled women and gender-diverse people.

    Call for mandatory gender impact assessments modelled on Canada's Gender-Based Analysis Plus, at the very least, for AI deployed in the public sector, with particular attention to intersectional harms affecting women of colour, women from lower socio- economic backgrounds, lower-income earners, those in precarious employment, migrant women, Indigenous women, women with disabilities, and gender-diverse persons.

  29. Legally require decision-specific AI explanations, contributing-variable information and accessible judicial/non-judicial contestation, consistent with the Council of Europe’s Equality and AI recommendation.

    Require meaningful transparency and explainability in law for AI-assisted decision- making, ensuring that individuals affected by automated decisions have legally enforceable rights to receive decision-specific explanations, to understand which variables contributed to the outcome and to contest adverse decisions through accessible judicial and non-judicial mechanisms, consistent with the Council of Europe's Recommendation on Equality and Artificial Intelligence (2026).

  30. Require continuous, gender-disaggregated post-deployment monitoring for AI systems in sensitive sectors, with deployers collecting and publishing performance data disaggregated by gender and intersecting characteristics.

    Require the establishment of continuous, gender-disaggregated post-deployment monitoring obligations for AI systems in sensitive sectors, requiring deployers to collect and publish performance data disaggregated by gender and intersecting characteristics.

  31. Amend the Online Safety Act 2021 to explicitly recognise the disproportionate impact of non-consensual deepfakes on women and girls.

    Recommend the amendment of the Online Safety Act 2021 to explicitly recognise the disproportionate impact of non-consensual deepfakes on women and girls.

  32. Ground AI regulation in CEDAW substantive equality: non-discriminatory access, meaningful transparency, informed decisions/consent, equal relations, vulnerability protection, effective remedies and gender-disaggregated monitoring.

    Urge the adoption of a substantive equality framework as the governing principle of AI regulation, grounded in Australia's obligations under the Convention on the Elimination of All Forms of Discrimination against Women (CEDAW) and require non-discriminatory access, meaningful transparency, informed decision-making and consent, promotion of equal relations, protection from vulnerability, effective remedies and gender- disaggregated monitoring to ensure that AI governance transforms, rather than reproduces, structural gender inequality.

  33. Ensure active participation of diverse women in AI governance, including gender quotas or targets on AI advisory bodies and regulators, funded engagement of women's organisations, and investment in closing AI literacy gaps.

    Ensure the active participation of diverse women in AI governance, including through gender quotas or targets on AI advisory bodies and regulatory institutions, funded engagement of women's organisations in regulatory design processes and investment in closing AI literacy gaps

  34. Embed regulatory frameworks for data centre development that recognise social, cultural and environmental impacts on local communities, with free, prior and informed consent and meaningful engagement.

    Embed regulatory frameworks concerning the development of data centres, recognising the social, cultural and environmental impacts these developments have on local communities. Particular attention needs to be paid to enabling free, prior and informed consent and meaningful engagement with communities.

  35. Have policymakers consider who is most exposed to AI harm, who shapes systems and who bears consequences when safeguards/accountability are inadequate.

    Policymakers should therefore consider who is most exposed to AI related harm, who is able to shape the systems being developed, and who ultimately bears the consequences when safeguards and accountability mechanisms are inadequate.

  36. Put forward Australian gender expertise and representation in national and international AI governance bodies.

    Ensure Australia puts forward gender expertise and representation in AI governance bodies (e.g. national taskforces or UN AI advisory bodies).

  37. Create a joint working group across foreign- and domestic-facing departments covering digital affairs/technology and gender equality/women’s issues.

    Create a joint working group across departments (foreign and domestic facing) relating to 1) digital affairs and technology and 2) gender equality and women’s issues.

  38. Fund gender-responsive AI innovation through grants, procurement and hackathons, prioritising gender-based violence, safety and women’s economic opportunities while encouraging co-design with gender experts and affected communities.

    Provide targeted innovation funding for gender-responsive AI tools (e.g. through grants, procurement tools, hackathon challenges). Investments should prioritise AI applications that address GBV, improve safety, and expand women's economic opportunities, while encouraging co-design with gender experts and affected communities.

  39. Require transparency in training datasets used for AI development, including authorship, content and sex- and gender-disaggregated data.

    Require transparency in training datasets used for AI development (e.g. attention to authorship, content, sex- and gender-disaggregated data).

  40. Fund tested and evidence-based AI tools for gender equality, such as gender-based violence prevention and safety systems.

    Fund tested and evidence-based AI tools to be used for gender equality e.g. around gender -based violence prevention, safety systems and so on.

  41. Adopt gender-responsive clauses in multilateral agreements on emerging technologies including AI, similar to gender action plans in climate agreements.

    Adopt gender-responsive clauses in multilateral agreements surrounding emerging technologies, including AI (similar to how climate agreements include gender action plans).

  42. Invest in research on localised gendered issues surrounding AI in Australia and partner countries, to develop evidence-based policy and practical interventions.

    Invest in research that seeks to understand localised gendered issues surrounding AI, in Australia and within partner countries we work, to develop evidence-based targeted options for policy and practical interventions.

  43. Bring together gender-equality leaders to inform multilateral AI governance, for instance through a Gender-Responsive AI Advisory Body.

    Bring together gender equality leaders to help inform multilateral AI governance, for instance through the establishment of a Gender-Responsive AI Advisory Body.

  44. Align Australian AI policy with UN Women, UNESCO and OECD frameworks, identifying implementation gaps and translating commitments into domestic guidance and accountability mechanisms.

    Align Australian AI policy with existing UN Women, UNESCO and OECD frameworks, identifying implementation gaps, and translating those commitments into domestic guidance and accountability mechanisms.

  45. Adopt human-rights-based, intersectional, inclusive, gender-transformative, victim-centred, trauma-informed, age-appropriate and child-friendly accountability for technology-facilitated violence against women and girls.

    We also echo the Council of Europe’s (2026) call for: a human rights-based, intersectional, inclusive, gender-transformative, victim-centred, trauma-informed, age-appropriate and child-friendly approach, to ensure accountability for all forms of technology-facilitated violence against women and girls.

Read the original

In some quotes a program corrected obvious text-recognition errors for display.

Open in the interactive guide