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 182

Pledge Human (Good Humans Good Work)

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.

Argues that AI policy should aim at human wellbeing and shared prosperity rather than productivity alone. Addresses who captures AI productivity gains, how workers and the public can influence AI decisions, and a proposed Public Interest AI Test. Makes three recommendations.

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 productivity should not be the objective of Australian AI policy; the objective should be human wellbeing and shared prosperity.

    But productivity itself should not be the objective of Australian AI policy. The objective should be improved human wellbeing and shared prosperity.

  2. Argues that who captures AI's productivity dividend depends on policy, markets, institutions and collective choices, not on technology alone.

    Who captures the productivity dividend? Technology does not determine the answer. Policy, markets, institutions and collective choices do.

  3. Argues that workers should have meaningful opportunities to influence significant technological changes to their work.

    Workers should have meaningful opportunities to influence significant technological changes to their work, particularly where those changes affect employment, surveillance, performance assessment or working conditions.

  4. Argues that AI governance cannot be separated from competition, labour, tax, social security, consumer protection and democratic governance.

    AI governance therefore cannot be separated from competition policy, labour policy, taxation, social security, consumer protection and democratic governance.

  5. Argues that government should help organisations solve problems rather than promote AI adoption for its own sake.

    Government should help organisations solve problems rather than encourage technology adoption for its own sake.

  6. Argues that when automated systems materially contribute to consequential decisions, people should be informed, able to challenge and given meaningful human review, with safeguards scaling with consequence.

    Where an automated system materially contributes to a consequential decision, people should know that it is being used, be able to challenge the decision, and have access to meaningful human review. The greater the consequence, the stronger the safeguard should be.

Recommendations

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

  1. Commission independent modelling of AI and robotics labour-market scenarios over 5, 10 and 20 years, including high-displacement scenarios, and investigate mechanisms for broadly distributing productivity gains.

    The Australian Government should commission independent modelling of AI and robotics labour-market scenarios over 5, 10 and 20 years, including high-displacement scenarios, and investigate mechanisms for broadly distributing productivity gains.

  2. Measure AI success beyond investment, adoption and GDP: wages, household income, hours, employment, job quality, labour–capital distribution, inequality, concentration, wellbeing and community outcomes.

    Australia should also measure AI’s success beyond investment, adoption and GDP, including: ● wages, household income and working hours; ● employment and job quality; ● distribution of productivity gains between labour and capital; ● inequality and economic concentration; and ● wellbeing and community outcomes.

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

  3. Develop mechanisms that give people meaningful influence over consequential uses of AI.

    Australia should develop mechanisms that give people meaningful influence over consequential uses of AI.

  4. In workplaces, include consultation before AI materially changes roles or working conditions, transparency where AI influences employment decisions, meaningful human review, and protections against unreasonable AI-enabled surveillance.

    For workplaces, this should include consultation before AI materially changes roles or working conditions; transparency where AI influences employment decisions; meaningful human review; and protections against unreasonable AI-enabled workplace surveillance.

  5. For broader technology policy, experiment with citizens' assemblies, deliberative panels involving affected workers and communities, and dedicated participation by young people, connected to decision-making.

    For broader technology policy, Australia should experiment with mechanisms such as citizens’ assemblies, deliberative panels involving affected workers and communities, and dedicated participation by young people. These processes should connect to decision-making rather than becoming consultation exercises with no clear consequence.

  6. Have AI literacy programs teach people not only how to use AI but about their rights and agency around it.

    AI literacy programs should similarly teach people not only how to use AI, but about their rights and agency around it.

  7. Help organisations solve problems rather than encourage technology adoption for its own sake.

    Government should help organisations solve problems rather than encourage technology adoption for its own sake.

  8. Where an automated system materially contributes to a consequential decision, ensure people know it is used, can challenge the decision and have meaningful human review, with stronger safeguards for greater consequence.

    Where an automated system materially contributes to a consequential decision, people should know that it is being used, be able to challenge the decision, and have access to meaningful human review. The greater the consequence, the stronger the safeguard should be.

  9. Apply a transparent Public Interest AI Test to significant Commonwealth AI policies, investments, incentives, procurement and high-impact deployments, examining benefits, costs, participation, workers, communities, economic concentration, failure, accountability and collective wellbeing.

    Significant Commonwealth AI policies, investments, incentives, procurement decisions and high-impact deployments should be assessed against a transparent Public Interest AI Test. It should ask: Who benefits? Who bears the costs? Who has a say? What happens to workers and communities? Does this increase or reduce concentrations of economic power? What happens if the technology fails? Who is accountable? And does this leave Australians collectively better off?

  10. For high-impact government AI, set proportionate requirements for transparency, impact assessment, human accountability, meaningful review and appeal.

    For high-impact government AI, this principle should translate into proportionate requirements for transparency, impact assessment, human accountability, meaningful review and appeal.

  11. Assess industry-policy economic benefits alongside employment, wealth distribution, competition, environmental costs, infrastructure requirements, sovereignty and dependence on individual technology providers.

    For industry policy, it should mean assessing economic benefits alongside employment, distribution of wealth, competition, environmental costs, infrastructure requirements, sovereignty and dependence on individual technology providers.

  12. Support small businesses to make good technology decisions rather than measuring success by AI adoption alone.

    And for small-business programs, it should mean supporting organisations to make good technology decisions rather than measuring success by AI adoption alone.

Read the original

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

Open in the interactive guide