Submission 287
Office of the Special Envoy to Combat Islamophobia (OSECI)
Submitter group: Government.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
Argues that AI systems can reproduce and amplify bias, with particular risks for Muslim Australians, across Commonwealth AI use, legal gaps on religious vilification, AI-generated hate content, and national security and foreign interference. Makes twelve recommendations including impact assessments, audits, transparency and recognising religion as a protected attribute.
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.
Argues that Commonwealth agency adoption of AI risks reproducing and legitimising existing prejudices through automated systems.
As Commonwealth agencies increasingly adopt AI to improve service delivery and administrative decision-making, there is a risk that existing prejudices may be reproduced, and legitimised, through automated systems.
Argues that Australia's legal frameworks were not designed for discrimination and vilification subject to algorithmic amplification, leaving significant gaps.
Australia’s legal and regulatory frameworks were not designed to address discrimination, vilification or hate crimes that are subject to algorithmic amplification. As a result, significant gaps exist in responding to the ways in which AI-driven recommender systems and platform algorithms systematically promote divisive, identity-based and discriminatory content.
Argues existing discrimination, vilification and hate crime laws do not distinguish platforms, developers, deployers and users, and need clearer shared responsibility.
Existing discrimination, vilification and hate crime laws do not differentiate between the actors involved: the online platform, the developer that trained the model, the business that deploys it, or the individual who enters the prompt. The law requires greater clarity on how responsibility is shared.
Argues that Muslims do not receive protection under section 18C and that there is no federal religious vilification protection.
Muslims are not considered a race or a single “ethno-religious group” and do not receive protection under section 18C of the Racial Discrimination Act 1975 (Cth), unlike Sikh and Jewish Australians. There is no federal protection against religious vilification
Argues that AI can become a force-multiplier of Islamophobia, enabling harmful narratives to spread faster and more persuasively.
AI can become a force-multiplier of Islamophobia, enabling harmful narratives to spread faster, further and more persuasively than before.
Argues that, without robust safeguards, data-driven systems risk embedding historical suspicion and deepening distrust.
Without robust safeguards, there is a significant risk that historical patterns of suspicion become embedded within data-driven systems, where they are harder to identify and challenge, deepening the trust deficit.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Require independent bias, fairness and human-rights impact assessments before high-risk AI systems are procured or deployed by Commonwealth agencies.
Require independent bias, fairness, and human-rights impact assessments before high-risk AI systems are procured or deployed by Commonwealth agencies.
Mandate regular auditing of AI systems, datasets and outputs to identify and address discriminatory effects, including religious and racial bias.
Mandate regular auditing of AI systems, datasets, and outputs to identify and address discriminatory effects, including religious and racial bias.
Require agencies to disclose the purpose, training-data sources, risk assessments and oversight mechanisms of AI systems used in public administration.
Require agencies to disclose the purpose, training-data sources, risk assessments and oversight mechanisms applicable to AI systems used in public administration.
Expand Australia's Christchurch Call commitments on algorithmic amplification of terrorist and violent extremist content to include discrimination, hate crimes and vilification.
As a founding supporter of the Christchurch Call, Australia has endorsed commitments to address how algorithms may drive users towards or amplify terrorist and violent extremist content.8 This commitment should be expanded further to include discrimination, hate crimes and vilification.
Adopt a multi-layered approach to legal responsibility for AI-generated harm applying to developers, deployers, designers and platforms, consistent with their ability to prevent harm at different stages.
The Committee adopt a multi-layered approach to legal responsibility for AI-generated harm that applies to developers, deployers, designers and platforms, consistent with their ability to prevent harm at different stages.
Recognise religion as a protected attribute within AI governance frameworks, including risk assessments, testing frameworks and algorithm design standards.
The Committee recognise religion as a protected attribute within AI governance frameworks, including in risk assessments, testing frameworks and algorithm design standards, to ensure the existing legal gap is not replicated.
Require developers of AI systems to conduct bias, discrimination and human rights impact assessments before public deployment, including religious, racial and gender-based stereotyping.
Require developers of AI systems to conduct bias, discrimination and human rights impact assessments prior to public deployment including religious, racial and gender-based stereotyping.
Establish mandatory transparency requirements for AI-generated content, such as watermarking or clear labelling, particularly where content targets vulnerable communities.
Establish mandatory transparency requirements for AI-generated content, such as watermarking or clear labelling, particularly where content targets vulnerable communities.
Invest in AI-enabled detection and early warning of emerging vilification trends, co-designed with affected communities.
Invest in AI-enabled detection and early warning of emerging vilification trends, co-designed with affected communities.
Establish a permanent community consultation mechanism within the AI governance framework so disproportionately affected communities are consulted in design, deployment and review of AI systems.
Establish a permanent community consultation mechanism within Australia’s AI governance framework to ensure disproportionately affected communities are meaningfully consulted in the design, deployment and ongoing review of AI systems.
Keep AI systems in law enforcement, immigration, counter-terrorism and national security subject to meaningful human oversight and annual independent auditing, so assessments are transparent, accountable and do not produce discriminatory outcomes for religious, ethnic or cultural communities.
Ensure AI systems deployed in law enforcement, immigration, counter-terrorism and national security remain subject to meaningful human oversight and annual independent auditing, so that automated assessments are transparent, accountable and do not produce discriminatory outcomes for religious, ethnic or cultural communities.
Expand foreign interference frameworks to explicitly recognise and address AI-generated misinformation designed to inflame social tensions targeting religious, racial or cultural communities.
Expand Australia’s foreign interference frameworks to explicitly recognise and address AI-generated misinformation designed to inflame social tensions targeting religious, racial or cultural communities.
Require digital platforms to establish rapid identification, labelling and removal mechanisms for AI-generated content linked to coordinated foreign interference campaigns, and report publicly on outcomes.
Require digital platforms to establish rapid identification, labelling and removal mechanisms for AI-generated content linked to coordinated foreign interference campaigns and report publicly on outcomes.