Submission 297
ARM Hub
Submitter group: Other.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
ARM Hub, a not-for-profit that operates a federally funded AI Adopt Centre, argues that AI lifts productivity only when firms adopt it, and that cost and implementation support are the main barriers for manufacturing SMEs. It asks for renewal of the AI Adopt Program, co-investment funding, a sovereign manufacturing dataset, and examination of a certification gap for machine-learning safety components.
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.
Argues that the gap is converting AI interest and experimentation into implemented business solutions.
The gap is conversion of AI interest and experimentation into implemented business solutions.
Says about 80 per cent of SMEs cite implementation cost, roughly $300,000 for advanced integration, as the barrier to deployment.
Second, cost, with no mechanism to de-risk it: about 80 per cent of SMEs cite implementation cost — roughly $300,000 for advanced integration — as the barrier to deployment
Argues that Australia's sovereign manufacturing data does not yet exist and that building the dataset requires urgent government attention.
Australia’s sovereign manufacturing data does not yet exist and needs to be created and collected. Building a sovereign manufacturing dataset is a once-only opportunity and requires urgent attention by the Australian Government.
Says evidence suggests AI commercialisation is best pursued outside universities, within a framework where universities are involved.
However, evidence suggests that the commercialisation function is best pursued outside universities, but within a framework where universities are involved in the process.
Says existing machinery safety rules suit deterministic components, but components whose safety behaviour comes from a trained model have no certification pathway.
Deterministic, sensor-based components — where behaviour is measured, not learned — fit cleanly within Australia's existing machinery safety regime (AS 4024, aligned to ISO 13849/IEC 62061) and can be certified through well-understood, self-assessed pathways. Components whose safety-relevant behaviour comes from a trained or foundation model have no such pathway.
Says the evidence does not support a mass-displacement narrative in the industrial sectors it works in, citing a 2025 meta-analysis of robotisation.
The evidence does not support a mass-displacement narrative in the industrial sectors we work in: a 2025 meta-analysis of 33 studies found the aggregate employment effect of robotisation to be minimal
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Provide ongoing co-investment funding matched by SMEs and released against verified milestones.
For around 80 per cent of SMEs, we have found cost is a major barrier, particularly for ARM Hub to deliver data and AI solutions that are built on a company’s own data. It is critical in this context that there is ongoing funding for co-investment, matched by participating SMEs and released against verified milestones.
Urgently create and collect sovereign manufacturing data for AI trained on Australian workflows and trade knowledge, recognising that no single firm can fund the dataset.
Australia’s sovereign manufacturing data does not yet exist and needs to be created and collected. Building a sovereign manufacturing dataset is a once-only opportunity and requires urgent attention by the Australian Government. AI models and tools for manufacturing production must be trained on Australian workflows and trade knowledge, and no single firm can fund building that dataset.
Take a national approach to shared sovereign data infrastructure, with Australian data residency, client ownership of data and the resulting work directed to Australian AI vendors and manufacturers.
This requires a national approach to shared sovereign data infrastructure – with Australian data residency, client ownership of data, and the resulting work directed for Australian AI vendors and manufacturers.
Renew the AI Adopt Program before it ends, with improvements such as re-weighting toward deployment and diffusion.
An immediate step would be renewal of the AI Adopt Program prior to the termination date with appropriate improvements as may be required, including re-weighting towards deployment and diffusion at the enterprise level.
Use renewal to recast business support programs more broadly into a more integrated approach to innovation and enterprise capability-building.
More fundamentally, renewal also provides the opportunity to recast business support programs more broadly. With a number of these programs cut or operating with reduced funding, the Government is well positioned to devise a more integrated approach to innovation and enterprise capability-building.
Pair regional engagement with enterprise-grade AI delivery capability and financial incentives, rather than another portal, workshop or toolkit.
Closing this gap requires regional engagement paired with enterprise-grade delivery capability, paired with financial incentive, not another portal, workshop or toolkit.
Look at place-based innovation ecosystems abroad (UK Catapult Centres, ManufacturingUSA Institutes, Canadian Clusters, Fraunhofer Network, Singapore A-Stars) for a proven research commercialisation framework.
Australia would benefit from looking at place-based innovation ecosystems around the world, such as the UK Catapult Centres, ManufacturingUSA Institutes, Canadian Clusters, Germany’s Fraunhofer Network and Singapore A-Stars, for a proven research commercialisation framework.
Work with existing and new industries to build the capabilities firms need to scale, including creating and adopting AI.
Australia must work with existing industries as well as new industries and build new capabilities that these companies need to scale their businesses, and this includes creating and adopting AI.
Calibrate matched implementation co-investment to the real cost of integration and tie it to verified milestones.
What remains unaddressed is implementation co-investment — matched funding calibrated to the real cost of integration and tied to verified milestones. It is the highest-leverage intervention available to government.
Use a coordinating intermediary to build critical mass for sovereign manufacturing data as shared Australian industrial infrastructure accessible across industry, rather than confined to one firm.
A coordinating intermediary is needed to generate critical mass and make the resulting dataset a genuine sovereign Australian asset — shared industrial infrastructure available across industry rather than locked inside one firm.
Address trade-knowledge loss and erosion of entry-level learning by capturing knowledge as data and pairing adoption with consultation, training and micro-credentials.
The real risks are the loss of undocumented trade knowledge as experienced workers retire, and the erosion of entry-level learning rungs in knowledge-work occupations. Both are design problems: capture trade knowledge as data, and pair adoption with consultation, training and micro-credentials.
Examine the gap in regulatory pathways to market for AI-based safety technology, which ARM Hub says is visible in ventures it works with today.
Both companies are building genuinely useful, deployable technology from the same accelerator cohort, addressing the same shared-workspace safety problem and only one of them has a clear regulatory pathway to market. That gap, visible in ventures ARM Hub works with today rather than hypothetical ones, is the practical case for this Committee to examine.
Avoid new compliance obligations falling disproportionately on SME manufacturers, and support practical guidance, standards alignment and assurance support instead.
We would caution against any new compliance obligations falling disproportionately on SME manufacturers, and would instead support practical guidance, standards alignment and assurance support of the kind the Centre at ARM Hub already provides.