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 313

Australian Technology Network of Universities

Submitter group: Industry bodies.

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 the barrier to AI benefits is organisational rather than technological and that Australia should fund AI diffusion and deployment. Recommends a national adoption network succeeding the AI Adopt Program, with applied universities (its members) as delivery partners, recognising deployment as fundable R&D activity, and a research compute set-aside.

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 the barriers to Australia gaining AI benefits while mitigating risks are organisational rather than technological.

    This environment means that the barriers to Australia gaining the benefits while mitigating the risks is organisational, not technological.

  2. Cites the RBA and Productivity Commission as identifying slow technology diffusion as a contributor to Australia's productivity slowdown.

    The Reserve Bank of Australia and the Productivity Commission have identified slow technology diffusion as a distinct contributor to Australia's productivity slowdown, predating the current AI cycle.

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

  3. Says open-weight models can be fine-tuned for Australian contexts and smaller models trained to fill domain-specific niches.

    Freely available open source (or “open-weight”) models can be fine-tuned for Australian contexts, and smaller models can be trained to fill domain-specific niches.

  4. Argues that, as with social housing in developments, incoming data centre investment should include domestic capabilities and public compute.

    Similar in principle to arrangements with developers to include some social or low cost housing as part of new developments, Australia can and should include domestic capabilities and public compute as part of the deal for incoming investment.

  5. Says it supports a training-data transparency requirement in forthcoming national standards.

    We support adding a transparency of training data requirement to forthcoming national standards, as per HTI’s submission to this Joint Select Committee.

  6. Argues the next dollar of AI investment returns more from building responsible, targeted capabilities than from training ever larger frontier models.

    This means that the next dollar of Australian AI investment returns more from building useful and responsible capabilities that target our particular goals that are at the frontier of sectors such as national security and health, rather than from training ever larger and more expensive general-purpose models like a new Claude or ChatGPT (i.e. “frontier” models).

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

Recommendations

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

  1. Have the Office of AI collaborate and draw on available expertise.

    We encourage the Office of AI to follow a collaborative approach and draw on the wealth of expertise available to them.

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

  2. Add a training-data transparency requirement to forthcoming national standards, as in HTI’s submission.

    We support adding a transparency of training data requirement to forthcoming national standards, as per HTI’s submission to this Joint Select Committee.

  3. Have the Office of AI and National AI Centre replace the AI Adopt Program after 2026–27 with an ongoing adoption network: multi-year applied-university/industry partners, SME management/workforce capability funding, and sustained-adoption evaluation.

    The Committee should recommend that the Office of AI, working with the National AI Centre, succeed the AI Adopt Program when it ends in 2026-27 with an ongoing national adoption network that: 18.1. funds delivery partners on multi-year terms, building on existing university and industry adoption structures with applied universities as delivery partners; 18.2. funds management and workforce capability in small and medium enterprises, not only technology uptake; and 18.3. is evaluated on sustained adoption by participating firms.

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

  4. Fund AI diffusion, adaptation and deployment through SERD Recommendation 6 for businesses outside the R&D Tax Incentive, and count industry-partnered deployment in the full-cost funding mechanism designed by the National Innovation Council and ATEC.

    The Committee should recommend that, in implementing Ambitious Australia, the Government recognise the diffusion, adaptation and deployment of AI as fundable activity by: 19.1. including it in the mechanism proposed in SERD Recommendation 6, which gives businesses outside the R&D Tax Incentive access to university research capability; and 19.2. counting industry-partnered deployment work in the full-cost research funding mechanism SERD Recommendation 2b asks the National Innovation Council and ATEC to design, so institutions are rewarded for demonstrated translation performance.

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

  5. Define the Office of AI's compute obligation as a set-aside proportion of capacity for research and innovation, with a published application pathway, allocation decisions and annual public reporting, located with the renting companies.

    The Committee should recommend that the Office of AI’s regulatory framework define the compute obligation as a Set-aside proportion of capacity to be made available for research and innovation, with a published application pathway, published allocation decisions and annual public reporting on the capacity provided. Since data centres can change ownership and the actual compute is controlled by the companies renting it, the obligation should be located with the renting companies (e.g. the major AI companies like Anthropic).

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

Read the original

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

Open in the interactive guide