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 109

Emeritus Professor Michael Quinlan and Associate Professor Penny Williams

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.

Open in the interactive guide

Summary

Read the original submission for the full text.

Two occupational health and safety researchers argue that algorithmic management and AI in workplaces bring some safety benefits but, in their reading of the evidence, harms that outweigh them, and that any AI legislation should align with OHS regimes. They review the EU AI Act's product-safety model, note workplace research is nascent, and recommend OHS regulatory reform.

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 algorithmic management has spread beyond gig and platform work into a wide range of industries.

    AM can be found a wide array of work settings covering almost if not all industries. This penetration is expanding and becoming more pervasive over time.

  2. Argues that whether a workplace system uses AI or algorithms may be opaque to workers and even to managers.

    AI or algorithms can be opaque. That is, whether a system uses AI or is under-pinned by algorithms may not be transparent to workers or even to the managers that use them.

  3. Describes a body of research linking algorithmic management to psychosocial hazards such as overload and stress from surveillance.

    Another body of research has linked AM to adverse effects on OHS, particularly psychosocial hazards such as overload and stress arising from surveillance.

  4. Says Australian OHS regulation gives limited recognition to algorithms, AI and digital work systems.

    At present there is limited recognition of algorithms/AI or the broader concept of digital work systems in Australian OHS regulation.

  5. Argues, from its analysis of the EU AI Act, that a product-safety approach alone cannot deliver safe deployment of management systems at work.

    As our analysis of the EU AI Act shows, the safe deployment of AM systems in the workplace cannot be achieved through product regulation alone.

  6. Says broad digital-work-systems definitions and authorised-officer access to algorithms should be emulated and extended in scope.

    The wide definition of digital work systems and the power of authorised officers to access algorithms in this legislation are important steps but need to emulate and extended in scope.

Recommendations

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

  1. Support stronger integration between AI governance and labour protection regimes, promote institutional oversight, and ensure meaningful worker participation in deploying and monitoring AI-based systems.

    Future policy development should support stronger integration between AI governance and labour protection regimes, promote institutional oversight, and ensure meaningful worker participation in the deployment and monitoring of AI-based systems.

  2. Ensure any overarching AI legislation either aligns with or provides scope for the different emphasis of OHS legislation, and recognise the limits of a single AI law resolving the issue.

    What we would like to stress or alert the committee to is regarding the passage of an overarching piece of AI legislation. There is clearly value in this, but we would indicate that such legislation needs to either align with or provide scope for the rather different emphasis in OHS legislation and the challenges if not limitations in seeing a single piece of AI legislation as resolving the issue.

  3. Have regulatory authorities provide clearer guidance on assessing algorithmic management’s workplace safety effects.

    It also requires that employers interpret their OSH duties to include the impact of AM systems and that regulatory authorities provide clearer guidance on how to conduct such assessments.

  4. Reconfigure OHS regulation to address algorithmic management, mandating regulator and union access to algorithms and encouraging negotiated industry-level codes of practice or guidelines.

    OSH regulation needs to be reconfigured to better address AM-related challenges including mandating regulator and union access to algorithms and encouraging negotiated industry level codes of practice or AM guidelines.

  5. Amend Australian OHS regulation for AI and make the NSW digital work systems approach national, avoiding complexity and interstate competition based on the lowest regulatory coverage.

    There is also a need to amend existing OHS regulation in Australia to better address the challenges of AI and provide avenues for synergies with AI legislation. The NSW Digital Work Systems Act is a worthy step towards this, but the shift needs to be made national so as to avoid complexity and costly forms of interstate competitions based on the lowest level of regulatory coverage.

  6. Emulate and extend the scope of broad digital work systems definitions and authorised-officer algorithm access.

    The wide definition of digital work systems and the power of authorised officers to access algorithms in this legislation are important steps but need to emulate and extended in scope.

  7. Make a more coordinated response by OHS regulatory agencies a priority, perhaps led by Safe Work Australia, with uniformity never lowering standards.

    As a priority, a more coordinated response by OHS regulatory agencies is needed, perhaps led by Safe Work Australia and with the strong caveat that uniformity should never entail a lowering of standards.

  8. Consider the Fair Work Commission as a venue for industry-wide negotiations on appropriate safeguards for introducing AI, alongside regulatory guidance from OHS regulators.

    Alongside regulatory guidance from OHS regulators, the Fair Work Commission could act as an venue for considering industry wide negotiations on appropriate safeguards regarding the introduction of AI.

Read the original

Open in the interactive guide