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 281

Australian Services Union (ASU)

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.

Open in the interactive guide

Summary

Read the original submission for the full text.

The Australian Services Union draws on 2025 and 2026 member surveys to describe AI's effects on jobs, surveillance, consultation, training and discrimination, while noting members' use of and interest in AI. It proposes a 'digital just transition' of five actions, including legislated consultation, a workplace AI Act, free public AI training, six months' notice for AI-related job loss, and shared productivity gains.

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 over 80% of surveyed members reported no formal workplace AI policy or did not know of one.

    Over 80% of members said their workplace either did not have a formal AI policy, or they didn’t know of one.

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

  2. Argues existing AI consultation obligations are inadequate and should be triggered before introduction and whenever system use or capability materially changes.

    Existing consultation obligations are not fit for purpose in relation to AI. Consultation obligations should be triggered before introduction and whenever the use or capability of an AI system materially changes.

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

  3. Argues that Privacy Act exemptions for employee data have allowed widespread workplace surveillance feeding AI or algorithms.

    Loopholes in the Privacy Act mean privacy obligations on those collecting personal information do not apply to employee data. As a result, widespread workplace surveillance has been allowed, meaning employers can collect vast swathes of information about workers, and feed this information through AI or algorithms.

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

  4. Argues that AI is set to replace many highly feminised jobs and could worsen gender segregation in the workforce.

    AI is set to replace many of the most highly feminised jobs, such as accounting, administration and customer service. AI could exacerbate pre-existing gender segregation in the workforce.

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

  5. Argues that AI-driven productivity gains could support shorter working hours without loss of pay, so workers share in the benefits.

    AI-driven productivity gains could support a 4-day working week and, over time, further reductions in standard working hours without loss of pay. Sharing productivity gains through reduced working time would ensure that workers benefit directly from technological change rather than bearing only its risks.

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

  6. Says employer-provided AI training is tied to the employer's program and non-transferable, and that a public short course would counter this.

    We see workers being trained by their employers using the employer’s specific program, meaning the skills they develop are non-transferable and relevant only to that program or workplace. A public and universal short course would help counter that.

Recommendations

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

  1. Adopt the ACTU submission's recommendations, including an AI licensing scheme (which the ASU endorses).

    The ASU also endorses the Australian Council of Trade Unions (ACTU) submission and their recommendations, including an AI licensing scheme.

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

  2. Trigger consultation before AI introduction and whenever its use or capability materially changes.

    Consultation obligations should be triggered before introduction and whenever the use or capability of an AI system materially changes.

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

  3. Guarantee water allocation for citizens before industrial needs.

    As a drought-prone nation, Australia must be considerate of its water usage, and guarantee water allocation for citizens before any industrial needs.

  4. Act proactively to prevent AI from increasing inequality.

    Given that there are already substantial wealth gaps in Australia, with the top 10% of households holding 46% of total wealth,? the Government must be proactive in ensuring inequality does not grow.

  5. Give each reform a clear implementation timeline for worker security and a mechanism for worker and union participation and review.

    Each reform should have a clear timeline for implementation to provide worker security, and a mechanism for worker and union participation and review.

  6. Guarantee workers universal, equitable access to AI and digital tools, enforceable before the Fair Work Commission.

    ASU members want fair and equal access to AI and digital tools. Workers need a universal, equitable and fair right to tools, guaranteed by employers and enforceable before the Fair Work Commission.

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

  7. Protect workers against being made to train AI for job replacement, penalise employers who do so, and adopt protections such as a Workplace Jobs Guarantee.

    Access to AI tools should be for the purpose of helping workers perform work, not for training AI systems to take over workers’ jobs. There must be protections for workers and penalties for employers who make workers train AI systems for job replacement. Protections within organisations like a Workplace Jobs Guarantee must be adopted.

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

  8. Work with unions and employers so employers guarantee workers access to digital and AI tools, with policies and procedures on how and when to use AI at work.

    The Federal Government must work with unions and employers to ensure employers guarantee access to digital and AI tools for workers, with corresponding policies and procedures for how and when to use AI at work.

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

  9. Roll out job-protecting AI access and training compacts economy-wide, with government helping organisations procure high-quality, ethical AI tools fairly and affordably.

    The ASU has been pursuing compacts with large national systems employers like Microsoft. Here, we have advocated for workers to have access to AI tools and training underscored by guarantees from Microsoft to protect jobs. This is a model that should be rolled out across the economy, actively pursued and facilitated by the Federal Government who should also assist organisations to access high-quality AI tools with ethical standards, procured fairly and affordably from AI developers.

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

  10. As part of AI compacts, encourage employers to take on young and entry-level workers to preserve career pathways.

    As part of these compacts, the Federal Government should also encourage employers to take on young workers, or workers in entry-level jobs to ensure future generations have career pathways in their chosen fields.

  11. Actively facilitate compacts between unions and employers for access to standardised digital tools embedded with workplace rights and protections.

    The Federal Government should actively facilitate compacts between unions and employers for access to standardised digital tools embedded with workplace rights and protections for workers.

  12. Have government, through bodies such as Jobs and Skills Australia, design and run free AI fundamentals courses through TAFEs or local councils, teaching universal basics.

    The Federal Government, through bodies such as Jobs and Skills Australia, should establish a free, short training course on AI fundamentals to be rolled out at TAFEs or delivered through local councils across the country. The Government should design and run the course so that training helps workers understand the universal basics of AI.

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

  13. Have the public AI course teach AI limitations and how to review and critically assess programs.

    A public and universal short course would help counter that. The course should also cover the limitations of AI and how to review and critically assess AI programs.

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

  14. Include data origins and storage, AI capabilities and uses, ethics and protection of vulnerable people, and hallucinations and misinformation in the universal public course.

    A public, universally provided AI training course should include: - The basics; what AI is, where data comes from and how it is stored, - Functionality; AI capabilities and how it can be used in different settings, with some examples from various widely used models like Chat GPT, Co-Pilot and Claude, - Ethics; when is it appropriate to use AI and how to protect clients and other vulnerable people, and - Alshortfalls; including hallucinations and mis/disinformation.

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

  15. Establish and run a standardised, free and universally accessible AI training course; employers to provide paid time off for all workers to attend.

    The Federal Government should establish and run a standardised, free and universally accessible AI training course. Employers should provide paid time off for all workers to attend training.

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

  16. Partner with community organisations to train and support jobseekers lacking technological proficiency; make the public AI course available to all jobseekers and pay them for training time.

    The Federal Government should partner with community organisations assisting jobseekers to provide detailed training and support for workers who are not technologically proficient. A publicly run short AI course should also be universally accessible to all jobseekers, who should be paid for their time in training.

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

  17. Support jobseekers to undertake training on digital tools and AI.

    The Federal Government should support jobseekers to undertake training on digital tools and AI.

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

  18. Legislate worker and union consultation before final decisions to introduce or materially alter AI affecting employment, duties, classifications, workload, allocation, rosters, pay, promotion, skills, privacy, surveillance, performance management, redundancy or job security.

    The ASU will be seeking to include consultation clauses in our agreements that require genuine and meaningful engagement with the union and the workforce before making a final decision to use AI, or materially alter the way AI is used. Consultation must occur with workers whose employment, duties, classifications, workload, work allocation, rostering, remuneration, promotion opportunities, skills, privacy, surveillance, performance management, redundancy or job security will be impacted. Similar consultation obligations should be reflected in the law.

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

  19. Legislate minimum employer consultation obligations with civil penalties for failure, reflected in regulations, with a model AI consultation clause for enterprise agreements.

    The Federal Government must legislate minimum consultation obligations for employers. There should be civil penalties where employers fail to meet their consultation obligations. Consultation obligations should also be reflected in regulations, with a specific model consultation clause for the introduction of AI to be incorporated into EAs.

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

  20. Embed unions’ rights to intervene and stop work for unacceptable AI harms, including physical or psychological safety, discrimination or suppressed union rights, and risks to vulnerable clients or the public.

    Worker voice and strong safeguards must include meaningful consultation rights. Unions must have the right to intervene and stop work where there is an unacceptable risk of harm due to AI, such as where: - There’s a high risk to health and safety, for both physical and psychological safety, - The use of AI would lead to discrimination against workers, including diminishing workers’ rights to join unions and undertake union advocacy, and/or - There is an unacceptable risk to vulnerable clients or the public.

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

  21. Codify employer obligations on collecting worker data, extending to storage and security, with workers able to access (and where available delete) data collected about them at any time.

    Obligations on employers collecting data about workers must be codified and should extend to data storage and security. Workers should have a right to access data collected about them (and, where available, to have it deleted) at any time.

  22. Ensure transparency about how worker data is reviewed and which AI systems are involved.

    There is no transparency about how this data is being reviewed and what AI systems might be involved in this process. This must be rectified.

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

  23. Include understanding data-collection programs and applying that knowledge to workplace rights within delegates’ paid AI and surveillance training.

    Delegates must also be trained in how to understand this data and the programs that are being used to collect data. Delegate AI and workplace surveillance training should form part of delegates’ paid right to training. Here, there should also be a focus on delegates being able to translate this information and apply it to understanding their workplace rights.

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

  24. Give delegates access to paid training to understand data collected about them and their co-workers.

    Delegates must be able to access paid training to understand data collected about them and their co-workers.

  25. Amend the National Employment Standards to provide at least six months' paid notice for workers who lose their jobs due to AI.

    The Federal Government should amend the National Employment Standards (NES) to provide paid notice of at least six months for workers who lose their jobs due to AI.

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

  26. Legislate a workplace AI Act establishing universal worker protections and enforceable duties across procurement, design, introduction, use, monitoring and review.

    The ASU joins with the ACTU in their call to legislate a new AI Act that establishes universal minimum standards and protections for workers as AI is rolled out across Australian workplaces. The Act should create enforceable rights and obligations across the full AI lifecycle, including procurement, design, introduction, use, monitoring and review.

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

  27. Require human review and a responsible manager for every AI-made industrial-relations decision; keep records enabling independent review under the agreement where AI materially informs an employment decision affecting an employee.

    There is a need for humans to review every Al- made decision in industrial relations, and for a human manager to be designated as responsible. Where AI materially informs an employment-related decision affecting an employee, the employer must keep sufficient records to enable the basis of the decision to be independently reviewed under this agreement.

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

  28. Strictly limit biometrics and inferencing technology, protecting human characteristics and workers’ right to join unions.

    Strict limitations on biometrics and inferencing technology: biometrics are automated identifications of unique human characteristics like fingerprints, facial recognition, or iris scans. ASU organisers report the use of facial scanning programs used alongside inferencing technologies intuit specific identifiers like mood. Overseas, this has been overlaid with programs that analyse likelihood of individual workers joining unions. There have also been reports of biometric data being collected for valid reasons, and then being poorly stored and hacked. The strongest protections are needed to ensure human characteristics are not put at risk and the right for workers to join their union is protected.

  29. Limit workplace surveillance to legitimate purposes such as necessary plant protection; enshrine access to collected information and notice of when and how surveillance occurs.

    Workplace surveillance should be strictly limited to legitimate purposes such as where it is necessary to protect plant. Workers should also have an enshrined right to access information collected about them and to be informed about when and how workplace surveillance is occurring.

  30. Update discrimination codes to address AI risks to vulnerable workers.

    Discrimination codes must be updated to adequately address the risks posed by AI.

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

  31. Introduce substantial fines for social media companies allowing misinformation to flourish on their platforms.

    The Government must introduce substantial fines for social media companies that allow misinformation to flourish on their platforms.

  32. Protect workers’ income when they cannot work because AI or algorithmic-management technology is unavailable.

    Employees should not be without pay if they cannot perform work because AI or AM technology cannot be used.

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

  33. Equip state OHS frameworks for emerging technological hazards; have the Federal Government work with States and Territories to ensure relevant OHS Acts protect workplace safety.

    State based OHS frameworks must be well equipped to deal with new hazards at work as new technologies evolve. The Federal Government should work with State and Territory governments to ensure the relevant OHS Acts protect workplace safety.

  34. Recognise and value essential, non-replicable human skills and develop recruitment, retention and remuneration strategies for sectors relying on them.

    The Federal Government must recognise and value essential, non-replicable human skills and develop recruitment, retention and remuneration strategies for sectors that rely on these skills.

  35. Recognise essential non-replicable skills and outline career pathways, using SCHADS Award recognition of lived experience as a model.

    These critical skills need to be recognised and pathways for career progression should be outlined. For the first time, lived experience has been recognised as a skill in the Social, Community, Home Care and Disability Services Industry (SCHADS) Award. This should serve as a model for the further recognition of other non- replicable essential skills.

  36. Allocate NDIS training funds accessible through paid time off, with one training credit per 50 hours worked for external accredited training, regardless of permanent, casual or contractor status.

    Specific funds need to be allocated to NDIS workforce training. Workers should be able to access this fund by accruing paid time off. The ASU proposes that NDIS workers accrue one training credit for every 50 hours of work, to be used to undertake external, accredited training. This means workers will be able to accrue training credits regardless of whether they are employed permanently, casually or as contractors.

  37. Provide immediate training for new NDIS recruits, Certificate III for workers lacking it, and ongoing accredited training through credits accumulated every 50 hours, including permanent, casual and sole-trader workers.

    The ASU’s plan for portable training entitlements for NDIS workers includes: - Immediate training for new NDIS recruits to provide adequate skills to start their careers in the NDIS, - Foundational training (Certificate III) for all NDIS workers who do not currently have this Qualification, and - Ongoing training and development for NDIS workers by accumulating training credits for every 50 hours worked in the NDIS. Whether working permanent, casual or as a sole trader, all NDIS workers would be able to accumulate training credits and access accredited courses.

  38. Implement long-term five-year funding cycles, with employers required to use the funding to create secure, ongoing jobs with good conditions.

    The Federal Government must implement long-term, five-year funding cycles with a corresponding commitment that employers will be required to ensure long-term funding is used to create secure, ongoing jobs with good conditions.

  39. Work with the ASU to provide NDIS workers a portable paid right to training, accruing a minimum of three days of accredited training per year on average.

    The Albanese Government should work with the ASU to provide a portable paid right to training for NDIS workers. On average, NDIS workers should be able to accrue entitlements and credits for a minimum of three days of accredited training per year.

  40. Legislate rights to work from home and work with unions to implement them as appropriate to workplaces and sectors.

    The Federal Government should legislate rights to work from home and work with unions to implement these rights as appropriate to workplaces and sectors.

  41. Include in work-from-home legislation a reverse onus requiring employers to prove legitimate business grounds for office attendance and give substantial notice before changing rights.

    Work-from-home legislation should include a reverse onus, under which employers must prove legitimate business grounds for requiring workers to come to the office and provide substantial notice before work-from-home rights are changed.

  42. Collaborate with employers, the ASU and Australian unions to implement the four-day work week across Australian workplaces.

    The Federal Government and employers should collaborate with the ASU and Australian unions to implement the four-day work week across Australian workplaces.

  43. Include organisations using AI and AI developers among the possible subjects of AI levies.

    How AI will be taxed also needs to be considered, especially if widespread redundancies erode the existing income tax base. Models such as levies on organisations that use AI, AI developers and data centres are worth considering.

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

  44. Consider extending workers’ image and likeness rights to collective ownership of information used or replicated by AI, and rights over workplace data and other information used to train or operate systems.

    International precedents also show that workers can secure rights over the commercial use of their image and likeness: in the United States, the actors’ union has won intellectual property rights for workers to own their image. This concept could be expanded to give workers collective ownership over information used and replicated by AI, and rights over workplace data and other information used to train or operate AI systems.

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

  45. Explore AI taxation models, including levies on data centres and on the use of inputs.

    The Federal Government should explore AI taxation models, including levies on data centres and on the use of inputs.

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

  46. Examine whether AI levy revenue could expand public housing, universal healthcare, and universal education and training, including AI training rights, public schools and tertiary institutions.

    The ASU recommends that the committee examine whether revenue raised through these mechanisms could support the expansion of essential services in the age of AI, including: - Public housing stocks, as workers impacted by AI might find it challenging to meet the costs of housing, - Universal health care, as health insurance and private health institutions might become out of reach for many impacted by AI and wealth inequality, - Universal education and training, including rights to AI training, and public schools and tertiary institutions,

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

  47. Examine whether AI levy revenue could support public transportation free or at significantly reduced fares.

    Public transportation, made free or with significantly reduced fares,

  48. Examine whether AI levy revenue could fund community-sector programs as support needs grow.

    Funding for community sector programs which will only grow in demand as more people need to access support,

  49. Examine whether AI levy revenue could support community programs providing essentials for dignity, including free technology and wi-fi access.

    Other community programs that ensure other items necessary to live with dignity, such as free access to technology and wi-fi, and

  50. Examine whether AI levy revenue could expand funding for council services such as libraries and for community-sector organisations.

    Expanded funding of services provided by councils like local libraries and by community sector organisations.

  51. Fund essential community services to ensure high-quality, universal and accessible provision of services needed for a good standard of living.

    Essential community services should be funded to ensure the high-quality, universal and accessible provision of services needed for a good standard of living.

Read the original

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

Open in the interactive guide