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 265

iSOFT Software Technologies

Submitter group: Companies.

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.

A profit-for-purpose technology company describes its Australian AI work and argues for recognising for-purpose organisations, of which it is one, in Commonwealth procurement, a practical definition of sovereign AI and a clearer pathway from Australian AI research to government adoption. It also supports a risk-based approach to AI regulation.

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. Recommends procurement preference or weighting for verified for-purpose organisations demonstrating measurable community, skills or Australian capability benefits, alongside value for money, competition, probity and performance.

    We recommend that the Government recognise verified for-purpose organisations within Commonwealth procurement by amending the Commonwealth Procurement Rules, or supporting procurement guidance, to provide appropriate preference or weighting where those organisations demonstrate measurable community, skills or Australian capability benefits. This should operate alongside - not instead of - value for money, competition, probity and performance.

  2. Argues sovereign AI should be defined by more than where data is hosted.

    Sovereign AI should mean more than simply hosting data in Australia. A practical definition should also consider who controls the data and intellectual property, whether systems can be audited, whether Australian teams can operate and assure them, and whether government can change providers without losing access to critical data, knowledge or capability.

  3. Argues Australian AI organisations should have a route from paid pilots to procurement and production.

    Australian organisations should have a practical opportunity to prove capability through competitive paid pilots, transparent evaluation and proportionate assurance. Where a pilot demonstrates value, there should be a credible pathway to procurement and production.

  4. Recommends risk-based AI regulation, with higher safeguards for material effects on pay, employment, eligibility, safety, sensitive information or public decisions, and lighter burdens for lower-risk productivity uses.

    Across these areas, we also support a risk-based approach to AI regulation. Higher safeguards should apply where AI materially affects pay, employment, eligibility, safety, sensitive information or public decision- making. Lower-risk productivity uses should not face the same burden.

  5. Describes its own approach of placing Australian rules first and human accountability throughout, with AI assistance third.

    Our approach is therefore to put authoritative Australian rules first, deterministic entitlement calculations second, AI assistance third, and human accountability throughout.

  6. States that it invests all of its profits in education, skills, local opportunity and community initiatives.

    100% of our profits are invested into education, skills, local opportunity and community initiatives that help more people participate in the technology economy.

Recommendations

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

  1. Recognise verified for-purpose organisations in Commonwealth procurement by amending the Procurement Rules or supporting guidance, with appropriate preference or weighting for demonstrated community, skills or Australian capability benefits, alongside and not instead of value for money, competition, probity and performance.

    We recommend that the Government recognise verified for-purpose organisations within Commonwealth procurement by amending the Commonwealth Procurement Rules, or supporting procurement guidance, to provide appropriate preference or weighting where those organisations demonstrate measurable community, skills or Australian capability benefits. This should operate alongside - not instead of - value for money, competition, probity and performance.

  2. Define sovereign AI beyond local hosting: consider data and IP control, auditability, Australian operational and assurance capability, and government provider switching without loss of critical data, knowledge or capability.

    Sovereign AI should mean more than simply hosting data in Australia. A practical definition should also consider who controls the data and intellectual property, whether systems can be audited, whether Australian teams can operate and assure them, and whether government can change providers without losing access to critical data, knowledge or capability.

  3. Offer Australian organisations competitive paid pilots, transparent evaluation and proportionate assurance, with a credible route to procurement and production where pilots demonstrate value.

    Australian organisations should have a practical opportunity to prove capability through competitive paid pilots, transparent evaluation and proportionate assurance. Where a pilot demonstrates value, there should be a credible pathway to procurement and production.

  4. Apply higher safeguards where AI materially affects pay, employment, eligibility, safety, sensitive information or public decisions; lighten lower-risk productivity burdens and retain human oversight, transparency, security and accountability in higher-impact settings.

    Across these areas, we also support a risk-based approach to AI regulation. Higher safeguards should apply where AI materially affects pay, employment, eligibility, safety, sensitive information or public decision- making. Lower-risk productivity uses should not face the same burden. In higher-impact settings, human oversight, transparency, security and clear accountability should remain central.

Read the original

Open in the interactive guide