Submission 175
SA Unions
Submitter group: Unions.
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.
South Australia's peak union council says AI could improve productivity and services but only if workers share the benefits. It argues for worker consultation, enforceable protections on privacy, human oversight, surveillance and intellectual property, a just transition, published agreements with AI and data-centre operators, and legislated data sovereignty.
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.
Says workers and communities should not bear AI's risks and costs while private companies capture its benefits.
SA Unions rejects any model in which workers and communities bear the risks and costs associated with AI while private companies capture the benefits.
Says consultation with workers must begin before AI systems are procured or implemented.
Genuine consultation must begin while options remain open and before employers procure, develop or implement AI systems.
Recommends prohibiting AI outputs as the sole justification for workplace decisions.
Employers must be prohibited from using AI system outputs as the sole justification for workplace decisions.
Says governments must publish agreements with AI companies and data centre operators, including incentives, concessions and infrastructure commitments.
Governments must publish agreements with AI companies and data centre operators, including financial incentives, tax concessions, discounted land, infrastructure commitments, preferential access to resources and other forms of assistance.
Says Australia must retain technical expertise, operational control and infrastructure to govern its data and maintain essential systems, not only server location.
Data sovereignty is more than just the physical location of a server. Australia must retain the technical expertise, operational control and infrastructure necessary to govern its data and maintain essential systems during cyber incidents, international disruption, or a failure or withdrawal of a private provider.
Recommends a just transition where AI materially changes or displaces existing work.
Where AI materially changes or displaces existing work, affected workers must be supported through a just transition.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Establish an enforceable regulatory framework centred on workers and the public interest, requiring private operators to bear their costs and sharing economic and productivity gains across the community.
Governments must establish a proactive and enforceable regulatory framework that places workers and the public interest at its core, requires private operators to bear the costs they create, and ensures that economic and productivity gains are shared across the community.
Combine sovereign infrastructure with robust data governance, security standards, access controls, independent auditing, incident response and a skilled workforce able to scrutinise systems.
Sovereign infrastructure must be accompanied by robust data governance, appropriate security standards, access controls, independent auditing, incident-response capability and a skilled workforce able to scrutinise the systems being used.
Judge government AI adoption by public-service accessibility, accuracy, responsiveness, accountability and community trust.
The adoption of AI by governments should ultimately be judged by whether it produces better public services. Relevant measures include accessibility, accuracy, responsiveness, accountability and community trust—not simply the number of systems deployed, the time reportedly saved or reductions in staffing.
Use AI to reduce administrative work, support professional judgement and free public servants for complex and public-facing work, alongside adequate staffing and training.
Service improvement cannot be achieved by treating technology as a substitute for an adequately staffed and properly trained public service. AI should instead reduce unnecessary administrative work, support professional judgement and give public servants more time to undertake complex and public-facing work.
Ensure workers receive AI's benefits through pay increases and shorter working weeks.
Workers must receive the benefits from the introduction of AI through a combination of pay increases and shorter working weeks.
Share AI and data-centre gains through secure jobs, better wages and conditions, skills, workforce capability, research, innovation and economic opportunities for affected communities.
The economic and productivity benefits of AI and data centres must be shared with workers and communities rather than captured solely by private interests. Benefits should include secure employment opportunities, improved wages and working conditions, investment in skills development, workforce capability building, research and innovation, and improved economic opportunities for affected communities.
Consult affected communities, particularly First Nations communities, and ensure they receive benefits proportionate to the impacts and private gains generated.
Affected communities, particularly First Nations communities, must be genuinely consulted and receive benefits proportionate to the impacts and private gains generated.
Give workers and unions meaningful information on proposed AI systems' purposes, data, decisions and likely effects on jobs, workloads, skills and conditions before procurement, development or implementation.
Genuine consultation must begin while options remain open and before employers procure, develop or implement AI systems. Workers and their unions must be provided with meaningful information about the purpose of a proposed system, the data it will use, the decisions it will influence and its likely effects on jobs, workloads, skills and working conditions.
Use AI to support workers, avoid unnecessary job losses, de-skilling and unreasonable workloads, improve job quality, remove repetitive or hazardous tasks and enable complex, skilled and interpersonal work.
AI should be used to support workers rather than to justify avoidable job losses, de-skilling or unreasonable work intensification. Time saved through automation should not automatically mean a higher workload or an expectation that workers continuously produce more with fewer colleagues. Technological change should instead improve job quality, remove repetitive or hazardous work and allow workers to devote greater attention to complex, skilled and interpersonal aspects of their roles.
Require employer-provided, paid, work-specific AI training and supervision covering limitations, appropriate uses, privacy, security, inaccurate or discriminatory outputs and when human review is required.
Workers must be provided with the information, training and supervision necessary to use AI systems safely and responsibly. Training should explain not only how to operate a system, but also its limitations, appropriate uses, privacy and security requirements, potential for inaccurate or discriminatory outputs, and the circumstances in which human review is required. It must be provided by employers, during paid working time, and tailored to the work being performed.
Where AI materially changes or displaces work, provide early notice, consultation, paid training, redeployment, transferable-skills recognition and adaptation resources; workers should not bear retraining costs.
Where AI materially changes or displaces existing work, affected workers must be supported through a just transition. This requires early notification and consultation, access to paid training, opportunities for redeployment, recognition of transferable skills and sufficient time and resources to adapt. Workers should not be required to carry the financial burden of retraining for changes introduced by their employer, nor should technological investment be used to abandon workers whose knowledge and labour helped make that investment possible.
Make workers and their representatives active, primary participants in AI adoption through consultation and collaboration at planning and implementation stages, before adoption.
Workers and their representatives must be active, primary participants in the adoption of AI technology through consultation and collaboration at both the planning and implementation stages, prior to AI adoption.
Provide workers affected by AI adoption with a fair transition: time, education and other resources to adjust, and preference for redeployment opportunities.
Workers who are affected by the adoption of technology including new ways of working through AI must be provided with a fair transition, where they are given time, education and other resources to adjust to the changing face of work and they are given preference for redeployment opportunities.
Prohibit employers from using AI outputs as the sole justification for workplace decisions.
Employers must be prohibited from using AI system outputs as the sole justification for workplace decisions.
Guarantee workers a right to privacy and intellectual property protection.
Workers must have a right to privacy, and intellectual property protection.
Give workers and representatives a right to an explanation when AI is used in decision making.
Workers and their representatives must have a “right to an explanation” when AI is used for decision making purposes.
Prohibit employers from using AI to make decisions affecting workers without human oversight.
And employers should be prohibited from using AI to make decisions affecting workers without "human oversight".
Ensure workers can challenge an AI-made decision or output without fear, retribution or other adverse consequences, and that AI systems without effective oversight are not used in managing workers, particularly performance management.
Where a decision or output has been made by an AI system, workers must be able to challenge it without fear, retribution or other adverse consequences. Furthermore, AI systems without effective oversight must not be used in the management of workers, particularly in performance management.
Do not use AI systems to make decisions about recruitment, promotion, performance evaluation and remuneration.
Consequently, AI systems should not be used to make decisions about recruitment, promotion, performance evaluation, and remuneration.
Consult workers and their representatives on data governance, data access, privacy and cybersecurity measures protecting AI systems and data storage.
Workers and their representatives must be consulted on the implementation of appropriate data governance, data retrieval and access, privacy and cybersecurity measures to appropriately protect AI systems and data storage.
Require employers to tell workers how their data is stored and protected.
Employers must be transparent with their workers as it relates to the ways that their data is stored, as well as how it is protected.
Require AI used for hiring and promotions to be free from bias, discrimination and exclusion of diverse workers.
AI used for hiring and promotions must be free from bias, discrimination and exclusion of diverse workers.
Involve workers and their representatives in developing guardrails so that gender pay gaps are not exacerbated and promotion opportunities are not limited through AI.
Workers and their representatives should be involved in developing guardrails to ensure that gender pay gaps are not exacerbated and promotion opportunities are not limited through the use of AI.
Require genuine, timely consultation on AI surveillance of workers, with workers able to opt out of surveillance not required for regulatory or work health and safety compliance.
Workers and their representatives must be informed and consulted on when, where and how AI is being used (or proposed to be introduced) by employers for surveillance of workers to establish transparency and guardrails. This consultation must be genuine, timely and allow workers to opt out of any surveillance not required for regulatory or work, health and safety compliance.
Prohibit surveillance or monitoring of communications between a trade union, its delegates, members and prospective members.
Employers must guarantee that surveillance or monitoring of communications between a trade union, its delegates, members and prospective members is prohibited.
Extend strong cybersecurity protection of private information to everyone interacting with an organisation, especially government agencies and related entities.
Employers have a duty to implement strong cybersecurity measures to safeguard workers' private information, and this obligation should also apply to anyone interacting with the organisation, especially when it is a government agency or related entity.
Strengthen copyright and intellectual property protections requiring creative workers' consent, attribution and fair remuneration, with specific Indigenous Cultural and Intellectual Property protections.
Governments must strengthen copyright and intellectual property protections to safeguard against the use of creative workers’ outputs without their consent, attribution and fair remuneration, with specific protections for Indigenous Cultural and Intellectual Property.
Subject government AI and data-centre support, procurement and investment to strong governance, accountability, oversight and transparency requirements serving the public interest.
Government support, procurement and investment in AI systems and data-centre infrastructure must be subject to strong governance, accountability, oversight and transparency requirements to ensure they are developed and operated in the public interest.
Publish all agreements with private firms, including incentives, concessions, infrastructure commitments and measurable public-benefit obligations.
All agreements with private firms must be published, including any incentives, concessions, infrastructure commitments and measurable public-benefit obligations.
Ensure commercial confidentiality does not prevent scrutiny of value provided, proponents’ obligations or government risk.
Claims of commercial confidentiality should not prevent public scrutiny of the value provided, the obligations accepted by proponents or the risk assumed by government.
Ensure governments can recover grants, concessions and other public support where a company fails to deliver promised employment, investment, infrastructure or community benefits.
Where a company fails to deliver promised employment, investment, infrastructure or community benefits, governments must be able to recover grants, concessions and other public support.
Make public-benefit obligations in agreements legally enforceable rather than voluntary aspirations.
Public-benefit obligations should be legally enforceable rather than treated as voluntary aspirations that disappear once a project has been approved.
Establish independent oversight of compliance with agreements, with workers, unions and affected communities having a meaningful role in monitoring delivery.
Independent oversight is therefore necessary to assess compliance and provide confidence that government decisions serve the public interest. Workers, unions and affected communities must have a meaningful role in monitoring whether commitments are being delivered.
Retain technical expertise, operational control and infrastructure to govern Australian data and maintain essential systems during cyber incidents, international disruption, or private-provider failure or withdrawal.
Australia must retain the technical expertise, operational control and infrastructure necessary to govern its data and maintain essential systems during cyber incidents, international disruption, or a failure or withdrawal of a private provider.
Incorporate cybersecurity, resilience and continuity requirements into government contracts, planning approvals and financial-support agreements, with sustained investment in the Australian workforce needed to oversee these systems.
Cybersecurity, resilience and continuity requirements must therefore be incorporated into government contracts, planning approvals and financial-support agreements. This must be accompanied by sustained investment in the Australian engineers, ICT professionals, cybersecurity specialists, data professionals and public servants required to oversee these systems and protect the information they contain.
Legislate data sovereignty and invest in sovereign AI and data capabilities protecting public-sector and government-held data from foreign interference, unauthorised access, manipulation and theft.
Governments must enshrine Australian data sovereignty in legislation and invest in sovereign AI and data capabilities that protect Australians’ data held by the public sector and government agencies from foreign interference, unauthorised access, manipulation and theft.
Invest in the technical expertise, professional workforce, professional standards and research capability needed for Australia to securely develop, operate and govern AI systems.
Governments must also invest in the technical expertise, professional workforce, professional standards and research capability required for Australia to securely develop, operate and govern AI systems.
Develop and operate AI infrastructure and data centres with cybersecurity, resilience and risk-management requirements protecting critical systems, Australian data and community interests.
AI infrastructure and data centres must be developed and operated with appropriate cybersecurity, resilience and risk management requirements to protect critical systems, Australian data and community interests.
Require operators seeking government approvals, incentives and support for AI infrastructure to demonstrate appropriate cybersecurity capability, workforce expertise and ongoing investment in skilled professionals.
Government approvals, incentives and support for AI infrastructure should require operators to demonstrate appropriate cybersecurity capability, workforce expertise and ongoing investment in the skilled professionals required to maintain secure and resilient systems.