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 151

Freelance Robotics Pty Ltd

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.

Freelance Robotics, a small Queensland robotics engineering firm that does feasibility scoping work, argues the main barrier to SME AI and automation adoption is the unfunded feasibility-scoping step. It also covers brownfield plant, integration skills, narrow sovereign AI models and grant design, with seven recommendations, and knows of no program measuring whether grant-funded automation stays in use two years after acquittal.

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 expensive, unfunded work to establish whether technology will suit a business, with no guaranteed usable answer, is the barrier to adoption.

    It is that the work required to find out whether a given technology will work in a given business is itself expensive, unfunded, and carries no guarantee of a usable answer.

  2. Says the engineering effort to bridge the gap on existing SME sites (retrofit, instrumentation, interfacing, guarding) is often a larger share of project cost than the AI or robot, and that grant programs are least willing to fund it.

    The engineering effort to bridge that gap — retrofit, instrumentation, interfacing, guarding — is frequently a larger share of project cost than the AI or the robot, and it is the part that grant programs are least willing to fund.

  3. Says the skills constraint is the scarcity of people in Australia who can specify, integrate, commission and maintain AI-enabled automated systems in a working plant.

    The constraint is that there are very few people in Australia who can specify, integrate, commission and then maintain an AI-enabled automated system in a working plant — and that an SME will never employ one.

  4. Argues that Australia need not compete in frontier general-purpose models because narrow, domain-specific models trained on unique data matter to Australian industry.

    Australia will not win a contest to build frontier general-purpose models, and it does not need to. The models that matter to Australian industry are narrow, domain-specific and trained on data that exists nowhere else:

  5. Argues that predictability of grant programs matters more than generosity for SME capital planning.

    Predictability matters more than generosity. An SME plans capital expenditure months ahead, and an automation project has a long lead time before a grant application is even possible — the scoping has to happen first.

Recommendations

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

  1. Establish nationally scaled, matched $10,000–$50,000 SME feasibility vouchers for independent on-site engineering assessment; fund answers, not equipment, count not proceeding as success, and build on NSW’s design at materially larger scale.

    Establish a national automation and AI feasibility voucher. A matched grant in the order of $10,000 to $50,000, available to SMEs to engage an independent engineering specialist to assess whether a proposed AI or automation investment is viable on their site — with the explicit condition that the money buys an answer, not equipment, and that a recommendation not to proceed is a successful outcome of the grant. The NSW Innovative Manufacturing Adoption Fund Stage 1 provides a design to build on, at a scale that would need to be materially larger to shift national adoption rates.

  2. Acquit feasibility vouchers on the quality of the investigation, not on whether a purchase followed.

    A feasibility voucher should be acquitted on the quality of the investigation, not on whether a purchase followed.

  3. Make existing-plant retrofit, instrumentation and integration eligible across Commonwealth and state manufacturing and technology-adoption programs, including sensors, interfaces, controls, safety systems and engineering labour to connect AI or robotics.

    Make retrofit, instrumentation and integration of existing plant expressly eligible expenditure. Across Commonwealth and state manufacturing and technology-adoption programs, eligible costs should include sensors, interfaces, control upgrades, safety systems and the engineering labour required to connect AI or robotics to machinery a business already owns — not only the purchase of new equipment.

  4. Fund a distinct civil advanced-manufacturing automation integration skills stream: apprenticeships, cadetships and mid-career pathways in controls, industrial vision, networking and functional safety through working integrators, beyond classroom-only delivery.

    Treat the automation integration workforce as distinct from general AI literacy, and fund it as such. Support apprenticeship, cadetship and mid-career conversion pathways specifically for automation and robotics integration — controls, industrial vision, industrial networking, functional safety — delivered through and alongside working integrators rather than only in classrooms. The mechanism for this already exists in Australia's apprenticeship and skills architecture; what is missing is a stream that names industrial automation integration as the target skill set for civil advanced manufacturing.

  5. Consider making a defined period of post-commissioning support and retraining an eligible or mandatory component of any grant-funded automation project.

    Consider making a defined period of post-commissioning support and retraining an eligible — or mandatory — component of any grant-funded automation project.

  6. Measure whether grant-funded automation remains in use two years after acquittal.

    we are not aware of any Australian program that measures whether grant-funded automation is still in use two years after acquittal. If the Committee makes only one evidence recommendation, we would suggest it be that.

  7. Recognise applied industrial AI as sovereign capability in program design; make CRC Projects AI and successor funding accessible to narrow-model engineering firms without presuming software-company applicants, research-institution leads or frontier-model uses.

    Recognise applied, domain-specific industrial AI as sovereign AI capability in program design. Ensure that AI-targeted funding — including the artificial intelligence allocations within the Cooperative Research Centres Projects program and successor initiatives — is reachable by engineering firms building narrow industrial models, and is not structured in a way that presumes a software-company applicant, a research-institution lead or a frontier-model use case.

  8. Publish and honour grant-round calendars where a cadence is promised, announce opening dates in advance and notify registered interest when dates change.

    Publish and hold to grant round calendars, and tier co-contribution by business size. Where a program commits to a round cadence, publish opening dates in advance and notify registered interest when they move.

  9. Scale grant co-contribution requirements to turnover or headcount.

    Scale co-contribution requirements to turnover or headcount.

Read the original

Open in the interactive guide