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 293

Victorian Trades Hall Council (VTHC)

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.

Argues employer self-regulation has failed and workers lack meaningful influence over AI, despite its productivity potential. Recommends worker voice and bargaining rights, an empowered whole-of-economy AI Safety Institute, high-risk workplace controls, prohibitions, enforceable penalties, Fair Work Act and surveillance reform, record-keeping, intellectual-property protection and higher redundancy pay.

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. Argues that employers use AI systems to manage and evaluate work, and make decisions affecting workers, without disclosure, consultation or consent.

    Every day, thousands of Victorian workers are impacted by their employers’ use of AI technology, automated decision-making (ADM) and surveillance. AI systems coordinate, manage and evaluate their work. Digital systems make vital decisions about their futures and their livelihoods without disclosure, consultation or consent.

  2. Argues that the findings of previous federal inquiries remain largely unimplemented despite the National AI Plan.

    Despite the Federal Government’s commendable efforts to ascertain a rich understanding of the risks and challenges posed by AI through several inquiries and the AI Plan, the findings of these inquiries remain, by and large, unimplemented.

  3. Argues existing consultation, workplace health and safety and privacy obligations have failed to prevent exploitative AI use by employers.

    Existing consultation, workplace health and safety and privacy obligations have failed to restrain employers from using AI in ways that entrench and intensify exploitation – often with little measurable productivity benefit.

  4. Describes a 2025 survey finding that 30% of Australians believe AI's benefits outweigh risks, the lowest of 47 surveyed countries.

    As of 2025, only 30% of Australians believe the benefits of AI outweigh the risks – the lowest ranking out of a 47-country survey with a sample size of 48,000.

  5. Argues AI, automation, surveillance and algorithmic management must be permitted in enterprise bargaining.

    AI, automation, surveillance and algorithmic management are increasingly fundamental characteristics of the modern workplace: playing an ever more significant role in determining the distribution of both power and profits between workers and management. These issues must be permitted as matters in enterprise bargaining.

  6. Says every previous inquiry it reviewed reached the same core findings: worker voice is essential, self-regulation is insufficient and consultation settings are likely inadequate.

    every single one of these inquiries was compelled to produce the same core assertions: worker voice is an indispensable asset, self-regulation is insufficient, the employment- context is fraught with AI-related risks, and existing workplace consultation settings are likely inadequate to meet the challenges of this technologies.

Recommendations

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

  1. Integrate worker voices at the company, industry and policy-making level.

    Governments must act now to integrate worker voices at the company, industry and policy- making level.

  2. Permit AI, automation, surveillance and algorithmic management as matters in enterprise bargaining, with workers able to future-proof agreements and unions free to bargain over the method, type and extent of automation.

    AI, automation, surveillance and algorithmic management are increasingly fundamental characteristics of the modern workplace: playing an ever more significant role in determining the distribution of both power and profits between workers and management. These issues must be permitted as matters in enterprise bargaining. Workers should also be free to ‘future-proof’ enterprise agreements from digitalisation impacts, as in Germany and the UK.41 Unions should be free to bargain over the method, type and extent of automation in workplace processes.

  3. Consider the additional ACTU and VTHC submissions to previous inquiries, which the recommendations are presented as complementary to.

    We present these recommendations as complementary to those presented by the ACTU (and VTHC) in previous inquiries on these matters, and encourage the committee to consider those additional submissions as it responds to the terms of reference.

  4. Elevate the roles, functions and powers of the AI Safety Institute so it can function as a fit-for-purpose, whole-of-economy regulator.

    Enhance the powers of the AI Safety Institute. Elevate the roles, functions and powers of the AI Safety Institute to the extent necessary for it to function as a fit-for-purpose, whole-of-economy regulator as recommended by previous Federal Inquiries.

  5. Give the AI Safety Institute EU AI Act-modelled enforcement powers to redesign AI risking public/workplace safety, collect information on high-risk systems' functions/workings and hold non-compliant organisations accountable.

    Ensure that the AI Safety Institute has the powers necessary to enforce the law and shape outcomes. In order to be effective, a broadened AI Safety Institute will need broad powers modelled on those provided for in the EU AI Act. The regulator must have the power to redesign AI systems where they pose a risk to public or workplace safety, collect information about the function and workings of high-risk systems and ensure accountability against organisations which are failing to comply with relevant obligations.

  6. Have independent audits performed against high-risk systems where deemed necessary.

    Compliance should not be automatically assumed, and independent audits should be performed against high-risk systems where deemed necessary.

  7. Give the regulator the ability to hear and respond to complaints or inquiries from unions, citizens, workers and community organisations about high-risk systems.

    The regulator should also have the ability to hear and respond to complaints or inquiries from unions, citizens, workers and community organisations about the workings or outcomes of high-risk systems.

  8. Legislate Institute-enforceable AI penalties deterring large firms; consider harm's nature/gravity/scale/intensity, offender size/turnover/market power, and how infringement became known.

    Legislate penalties against organisations which fail to comply with relevant obligations. The AI Safety Institute must be able to enforce penalties for violations of the law as it pertains to AI development or deployment, and these penalties should be sufficiently onerous to have a deterrent effect on large firms. Penalties should be calculated with due consideration of: i. The nature, gravity, scale and intensity of harm resulting from the infringement; ii. The size, annual turnover and market power of the offender; iii. The manner in which the infringement became known;

  9. Also consider infringement-related financial gains or avoided losses, including indirect benefits; industrial-relations benefits; cooperation with supervisory authorities, public-interest and registered employee organisations; and intentional or negligent conduct.

    iv. Any financial benefits gained, or financial losses avoided, including indirectly, from the infringement; v. Any industrial relations benefits gained for the offender related to the infringement; vi. The offender’s degree of co-operation with supervisory authorities, public interest organisations, and registered employee organisations; and vii. The intentional or negligent character of the infringement.

  10. Immediately establish a review of the Fair Work Act 2009 as it relates to AI, to clarify ambiguities about consultation, bargaining and industrial disputes.

    Immediately establish a review of the Fair Work Act 2009 (Cth) as it relates to AI. Ambiguities about how AI relates to consultation, bargaining, industrial disputes urgently need to be clarified. These regulatory uncertainties are being widely exploited by employers to erode workers’ fundamental rights, including the right to organise, workplace privacy, job security, and adequate rest.

  11. Review the Fair Work Act 2009 to ensure decision making using AI and ADM is covered and employers remain liable for these decisions.

    the Federal Government should review the Fair Work Act 2009 (Cth) to ensure decision making using AI and ADM is covered under the Act, and employers remain liable for these decisions.

  12. Coordinate state and territory workplace-surveillance regulation in consultation with unions to protect reasonable privacy, safety and industrial rights.

    Coordinate state and territory level efforts to guarantee reasonable workplace privacy and regulate workplace surveillance. Longstanding issues with the Privacy Act’s employee records exemption have become even more severe since the onset of AI. As recommended previously, the Federal Government should encourage and coordinate efforts to regulate workplace surveillance across Australia with a view to protecting workplace privacy, safety and industrial rights. These efforts should be undertaken in consultation with unions.

  13. Mandate oversight and ACTU/VTHC-modelled guardrails for high-risk AI; classify employment uses as high-risk, including recruitment, referral, hiring, remuneration, promotion, training, apprenticeship, transfer and termination.

    Subject high-risk AI systems to mandatory oversights and controls and classify AI systems used for employment related purposes as high-risk. This should include AI systems used for recruitment, referral, hiring, remuneration, promotion, training, apprenticeship, transfer or termination. Controls on these systems should be modelled of the guardrails recommended by the ACTU and VTHC in their respective responses to the DISR Guardrails Position Paper.

  14. Ban development, deployment, sale, uptake, distribution and export of AI intended to calculate workers' emotions in employment, civic, public or educational contexts or sort workers by protected attributes, including union membership.

    The Federal Government should legislate prohibitions on the use of AI where it would pose an unacceptable risk to workers, consumers and the community. Prohibitions here would protect Australians from harm will improving the social license and public trust in the AI rollout. There should be a total ban on the development, deployment, sale, uptake, distribution or export of AI systems that are intended to: i. Calculate the emotions of a worker in an employment, civic, public, or educational context; ii. Calculate, categorise, or sort workers according to protected personal attributes, nothing that trade union membership should be considered a protected attribute;

  15. Also ban AI intended to defraud/cause substantial economic losses or collect worker/performance data for AI limiting workers' informed decisions or changing behaviour to harm themselves, others or the community.

    iii. Defraud the public or cause substantial economic losses to groups or individuals; iv. Collect data or information about workers or work performance to develop or deploy an AI system that limits their ability to make informed decisions or changes their behaviour to harm themselves, others or the community;

  16. Ban mass remote biometric identification in public spaces, real-time or retrospective, with reasonable, proportionate exceptions for critical national-security, defence, counterterrorism and public-order cases.

    v. Perform real-time or retrospective mass remote biometric identification in public spaces. Reasonable and proportionate exceptions to this prohibition should be included for critical instances involving national security, national defence, counterterrorism and other matters pertaining to public order.

  17. Require employers' AI records to detail severity/frequency of worker challenges, effects on safety, rights and psychosocial conditions, and harm incidents; give regulators access, including breaches.

    Codify AI record keeping requirements. Ensure that employers using AI are required to keep detailed information about the severity and frequency of challenges being made to the system by workers; how the system impacts occupational health and safety, workers’ rights, and the psychosocial work environment; and any incidents which cause harm to workers or the community. Regulators must have access to these records including breaches.

  18. Legislate protection for Australians' collective intellectual/creative property; consider how best to protect workers' knowledge, property and cultural outputs from AI-development theft, especially First Nations intellectual/artistic property.

    Legislate to protect Australians’ collective intellectual and creative property. The Federal Government must ensure that consideration is given to how best to protect the collective knowledge, property and cultural outputs of Australian workers from being stolen for use in the development of AI systems. This is especially important with regards to cultural theft from First Nations intellectual and artistic property.

  19. Increase redundancy pay for workers in cases of AI displacement.

    Increase redundancy pay for workers in cases of AI displacement. The total sum of redundancy pay for workers should be increased in response to the risk of growing labour displacement.

  20. Prohibit covert AI performance monitoring; ensure workers/representatives know what is measured, can contest findings and identify the liable human; immediately inform them of procedural unfairness, discrimination or unjustified adverse action.

    Ban AI from covertly evaluating employee performance. The Federal Government should prohibit undisclosed AI-driven performance monitoring in the workplace. When an AI system is used to make, inform or evaluate decisions about worker performance, workers and their representatives must have a right to know what these systems are measuring, how to contest their findings, and what human is liable for their outcomes. Workers and their representatives must be immediately informed if such a system violates procedural fairness, contributes to discrimination or results in unjustified adverse action against a worker.

Read the original

Open in the interactive guide