Submission 64
Humanware Institute
Submitter group: Research institutes.
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.
Argues Australia should build a 'Humanware' layer, the human judgement and accountability that directs AI, alongside hardware and software. Proposes five actions: prioritising Humanware, building capacity to judge AI, national missions, a critical AI dependency test, and a public AI Transition Monitor. Discloses that its Project Artesian programme is seeking participation.
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.
Describes Humanware as the human layer determining technology’s purposes, direction and accountability.
Humanware is the human layer that determines what technology is used for, how it is directed and who answers for the result.
Argues that fast adoption alone is not enough, as it can make existing work quicker without improving the work itself.
Fast adoption alone is not enough. It can make existing work quicker without improving the work itself. It can also create hidden dependencies, weaken expertise and make accountability harder to locate.
Recommends making judging, directing, challenging and taking responsibility for AI common outcomes across education, vocational training, professional standards and the Australian Public Service.
Make the ability to judge, direct, challenge and take responsibility for AI a common outcome across education, vocational training, professional standards and the Australian Public Service.
Recommends three time-bound missions addressing practical AI transition problems in critical systems such as finance, health or energy.
Commission three time-bound national missions that solve a practical AI transition problem in a critical system such as finance, health or energy.
Proposes testing significant Commonwealth AI investments and critical systems for dependencies on models, cloud, compute, data, vendors and standards.
Apply a Critical AI Dependency Test to significant Commonwealth AI investments and to critical systems, so Australia can see where it relies on particular models, cloud and compute infrastructure, data, vendors or technical standards.
Proposes a public AI Transition Monitor updated continuously where evidence allows and when material change occurs, beyond periodic reporting cycles.
Create a publicly accessible AI Transition Monitor that is updated continuously where evidence allows and when material change occurs, not only on monthly, quarterly or annual reporting cycles.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Place Humanware alongside hardware and software in the national AI effort and assign one minister and one office responsibility.
Place Humanware alongside hardware and software in Australia’s national AI effort, and give one minister and one office clear responsibility for developing it.
Within 100 days, publish a national Humanware outcome and name the accountable minister and office.
Within 100 days, publish a short national Humanware outcome and name the accountable minister and office.
Map existing Commonwealth AI activity against what it builds for people, organisations and Australia.
Map existing Commonwealth AI activity against what it builds for people, organisations and Australia.
Require significant Commonwealth AI initiatives to explain how they will strengthen judgement, accountability and national choice.
Require significant Commonwealth AI initiatives to explain how they will strengthen judgement, accountability and national choice.
Make judging, directing, challenging and taking responsibility for AI common outcomes across education, vocational training, professional standards and the Australian Public Service.
Make the ability to judge, direct, challenge and take responsibility for AI a common outcome across education, vocational training, professional standards and the Australian Public Service.
Agree common Humanware capabilities and map them into existing education, VET, professional and public-service pathways.
Agree a small set of common Humanware capabilities and map them into existing education, VET, professional and public-service pathways.
Start capability-building with consequential decision-makers, including public servants, boards, regulated professionals, educators and frontline workers.
Start with people making consequential decisions, including public servants, boards, regulated professionals, educators and frontline workers.
Update learning examples and priorities continually as technology and practice change, without waiting for whole qualifications to be rewritten.
Update learning examples and priorities continually as technology and practice change, without waiting for entire qualifications to be rewritten.
Commission three time-bound national missions solving practical AI transition problems in critical systems such as finance, health or energy.
Commission three time-bound national missions that solve a practical AI transition problem in a critical system such as finance, health or energy.
Select three mission problems for consequence, readiness and public value, rather than a preferred technology or supplier.
Select three problems on the basis of consequence, readiness and public value, not a preferred technology or supplier.
Bring government, regulators, industry, researchers, workers and affected people together around a clear public outcome.
Bring together government, regulators, industry, researchers, workers and affected people around a clear public outcome.
Test system operation, refusal, failure, disputes and attacks; transfer proven changes to an organisation able to maintain them.
Test what should happen when systems work, refuse an instruction, fail, are disputed or are attacked, then transfer any proven change to an organisation able to maintain it.
Apply a Critical AI Dependency Test to significant Commonwealth AI investments and critical systems.
Apply a Critical AI Dependency Test to significant Commonwealth AI investments and to critical systems, so Australia can see where it relies on particular models, cloud and compute infrastructure, data, vendors or technical standards.
Map highest-consequence Commonwealth and critical-system AI dependencies.
Map the highest-consequence dependencies across Commonwealth and critical-system AI.
Require major procurement and system decisions to address evidence access, interoperability, continuity, switching provider and exiting.
Require major procurement and system decisions to address evidence access, interoperability, continuity, switching provider and exiting.
Test whether essential services can continue through failure, withdrawal, changed commercial terms or geopolitical disruption.
Test whether essential services can continue through failure, withdrawal, changed commercial terms or geopolitical disruption.
Create a public AI Transition Monitor updated continuously where evidence allows and when material change occurs, beyond periodic reporting cycles.
Create a publicly accessible AI Transition Monitor that is updated continuously where evidence allows and when material change occurs, not only on monthly, quarterly or annual reporting cycles.
Build the Monitor on existing government, regulatory, industry and research evidence, rather than a new data system.
Build on existing government, regulatory, industry and research evidence rather than create a new data system from scratch.
Show AI changes to work and essential services, benefits reaching regions and small and family businesses, emerging capability, harms and dependencies, and action taken.
Show where AI is changing work and essential services; whether benefits are reaching regional Australia, small businesses and family businesses; what capability, harms and dependencies are emerging; and what action is being taken.
Use current or event-driven signals where possible, publish uncertainty and make material change trigger a named decision or response.
Use current or event-driven signals where possible, publish uncertainty, and make material change trigger a named decision or response.