Submission 405
Dr Larry Stillman
Submitter group: Individuals.
Position labels are not shown for individual submitters.
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 ethical, social and environmental capability must be part of Australia's AI capability. Proposes social-good principles in generative AI regulation, renewable energy requirements for data centres, a national statement on ethical conduct in generative AI research, university ethics review and training, and civil society representation in AI policy.
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 ethical, social and environmental capability must form part of Australia's AI capability.
The central argument is that ethical, social and environmental capability must form part of Australia’s AI capability.
Argues generative AI is ethically acceptable only when it expands human judgement, protects those most exposed to harm and distributes benefits justly.
Generative AI is ethically acceptable only when it expands, rather than substitutes for, human judgement, social relationships, community capability and democratic control; protects those most exposed to harm; and distributes benefits, knowledge and power justly.
Distinguishes risk from misalignment of AI systems, which it says is difficult but in principle preventable through rigorous design, testing, safeguards and governance.
One class of risk arises from misalignment: a Generative AI system may behave in ways that depart from human intentions or control. Such risks are difficult but, in principle, preventable through rigorous design, testing, safeguards and governance.
Argues the environmental and physical impacts of generative AI must be dealt with promptly, and that data centres should be required to use renewable energy.
The environmental and physical impacts of Generative AI (such as placement, noise, power and water consumption) must be dealt with in a timely and responsible manner in an era of global warming. A requirement to use renewable energy by data centres is obvious.
Argues university generative AI research should be submitted to ethics offices, with ongoing external evaluation.
University and related entity research and training, undergraduate, Masters, PhD and staff research should be required to submit Generative AI R & D applications to their Ethics offices. Because of the pace of development and danger of the unknown, there should be ongoing external evaluation to ensure that ethical requirements are being met.
Argues that civil society, especially organisations representing those most likely to be excluded or harmed, should have a place in AI policy.
As well as this, civil society—especially organisations representing those most likely to be excluded or harmed—must have a place at the table if Generative AI is not to deepen social division and weaken social cohesion.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Subject the legislative framework and regulations covering generative AI providers to an overarching principle of social good (for humans and the planet) and trustable systems, with international cooperation.
The legislative framework and regulations that cover Generative AI providers by Australia should be subject to the overarching principle of social good (for humans and the planet), and trustable systems. The problem is global and there should be cooperation with others.
Deal in a timely and responsible way with the environmental and physical impacts of generative AI (placement, noise, power, water), including a requirement that data centres use renewable energy.
The environmental and physical impacts of Generative AI (such as placement, noise, power and water consumption) must be dealt with in a timely and responsible manner in an era of global warming. A requirement to use renewable energy by data centres is obvious.
Develop a National Statement on Ethical Conduct in Generative AI related Research and Development for Humans and Global Concerns, modelled on the 2025 human-research statement, for universities and other research institutions and entities.
A National Statement on Ethical Conduct in Generative AI related Research and Development for Humans and Global Concerns should be developed along the lines of the National Statement on Ethical Conduct in Human Research (2025) for universities and other research institutions and entities.
Require university and related entity research and training to submit generative AI R&D applications to ethics offices, with ongoing external evaluation to ensure ethical requirements are met.
University and related entity research and training, undergraduate, Masters, PhD and staff research should be required to submit Generative AI R & D applications to their Ethics offices. Because of the pace of development and danger of the unknown, there should be ongoing external evaluation to ensure that ethical requirements are being met.
Require universities and other research entities to develop ongoing training for staff and students in generative AI learning, research and development, including mandatory modules in all generative AI courses.
Universities and other research entities be required to develop ongoing training for staff and students involved in Generative AI related Learning, Research and Development for Humans and Global concerns. This should include mandatory modules in all Generative AI courses.
Give civil society, especially organisations representing those most likely to be excluded or harmed, representation in generative AI policy.
As well as this, civil society—especially organisations representing those most likely to be excluded or harmed—must have a place at the table if Generative AI is not to deepen social division and weaken social cohesion.
Support universities, through government, to develop effective teaching and culture change solutions for making ethical thinking part of AI development.
There is a challenge making ethical thinking a formal part of AI development. Universities should be supported by government to develop effective teaching and culture change solutions.
For power and control, assess research merit and identify control, burdens, benefits and community interests.
Community voice, ownership, effective use Common good; persons are Proprietary concentration, capability diffusion Research merit; identify control, burden, benefit Power and control and democratic accountability. not instruments. and unaccountable deployment. and community interest.
For privacy and surveillance, provide privacy, confidentiality, cultural sensitivity and a data-governance plan.
Privacy and Resist centralisation, data extraction and Dignity and freedom from Mass profiling, identity resolution, and Privacy, confidentiality, cultural sensitivity and a surveillance function creep. intrusive control. transcript or training-data exposure. data-governance plan.
For vulnerability and inequality, provide risk-sensitive accessible consent, protection and empowerment.
Work with, not on, marginalised groups; Vulnerability and Care and empowerment of AI divide; low skill thresholds for abusers; Risk-sensitive, accessible consent; protection build agency through trusted inequality those at risk of exclusion. unequal ability to recognise or contest harm. and empowerment. intermediaries.
For participation, culture and inclusion, require consultation and respect for beliefs, customs, heritage and agreements.
Participation, culture Co-design; plural knowledge; language Participation, subsidiarity and Cultural flattening, dominant-language bias Consultation; respect for beliefs, customs, and inclusion and cultural authority. human diversity. and exclusion from design. heritage and agreements.
For knowledge, ownership and commerce, establish clear IP, data, commercial-interest and benefit-sharing arrangements.
Knowledge, Knowledge and infrastructure are Economic activity serves the Commodification of data and creative work; Clear IP, data, commercial-interest and benefit- ownership and community assets and public goods. common good. model theft and unauthorised reuse. sharing arrangements. commerce
For error, opacity and bias, require sound methods, integrity, disclosure of limits, fair treatment and appeal.
Error, opacity and Scrutiny, translation and contestability of Truth and responsible human Hallucination, biased classification, deceptive Sound methods, integrity, disclosure of limits, fair bias technical claims. judgement. behaviour and poor evaluation. treatment and appeal.
For systemic truthfulness, require honesty, transparency and research integrity; disclose AI’s role and limits, verify outputs and sources, protect against deception, foreseeable misinformation and fraud, and provide accessible correction and complaint avenues.
Communities need trustworthy information, Provenance and disclosure of AI material; Honesty, transparency and research integrity; Truthfulness, conscience, understandable explanations and the prevention of deceptive interfaces, fabricated disclose AI’s role and limits; verify outputs and Systemic dialogue and social trust; capacity to question, correct and contest evidence, deepfakes, impersonation and AI- sources; protect against deception, foreseeable truthfulness communication serves dignity systems shaping public knowledge and enabled fraud; audit trails and correction misinformation and fraud; provide accessible and the common good. decisions. mechanisms. correction and complaint avenues.
For misinformation and democracy, assess community harm, communicate findings responsibly and monitor risk.
Misinformation and Strengthen civic intelligence, dialogue and Truth, dialogue and social Deepfakes, automated personas, targeted Assess community harm; communicate findings democracy informed self-governance. trust. propaganda and influence operations. responsibly and monitor risk.
For education, work and care, protect autonomy and welfare and minimise psychological, social and economic harms.
Education, work and Build critical capability; judge social Dignity of work, learning and Deskilling, displacement, automated Protect autonomy and welfare; minimise care outcomes, not efficiency alone. care. monitoring and weakened care relationships. psychological, social and economic harms.
For war, peace and security, require dual-use assessment, restricted access and human accountability; stop if risk outweighs benefit.
Peace, justice, non- Cyber operations, weapons design, War, peace and Reject security systems that normalise Dual-use assessment; restricted access; human dehumanisation and protection autonomous targeting and harmful security surveillance, classification and exclusion. accountability; stop if risk outweighs benefit. of life. compliance.
For environmental impacts, require proportionate benefit, lower-impact methods and consideration of affected communities.
Energy and water demand, mineral Environmental Technology is material infrastructure, Care for creation and future Proportionate benefit; lower-impact methods; extraction, data centres, e-waste and rapid impacts affecting communities and the biosphere. generations. consider affected communities. scaling.
For ongoing accountability, retain researcher and institutional responsibility: monitor harms, report incidents, review risk, and modify or stop research when risks are no longer justified.
Researchers and institutions retain responsibility: Responsibility, prudence and Misuse, misalignment, operational failure and Ongoing Governance and solidarity are continuous, monitor harms, report incidents, review risk, and moral accountability cannot be loss of control require monitoring, audit, accountability not one-off technical fixes. modify or stop research where risks are no delegated to a machine. incident reporting and review. longer justified.
Support the social-services workforce.
support for the social-services workforce.
Align the proposed national statement internationally.
Proposes an Australian National Statement on Ethical Conduct in Generative AI Research and 9. Laws and regulatory frameworks Very strong overlap. Development, modelled on the National Statement on Ethical Conduct in Human Research, and aligned internationally.
Develop data governance that respects personal, cultural and community-held knowledge.
Raises the need for data governance that respects personal, cultural and community-held knowledge.