From the Research Universities edition of October 22, 2026
Science & Technology Australia National AI in Research Survey (1,034 responses): about three-quarters of researchers use AI, 54% call institutional training inadequate, and the ~10% who sized compute needs total 53.5M CPU-hours and 14.3M GPU-hours
Science & Technology Australia (STA), with MetaMelb at the University of Melbourne · Research enterprise (higher education, public research agencies, medical research institutes) · Australia (international; methods mirror a UK survey)
- Publisher
- Science & Technology Australia
- Original publication
- 20 October 2026 (STA media release, issued as STA gave evidence to the Joint Select Committee on Artificial Intelligence)
- Source retrieved
- 2026-10-23
- Event date
- 2026-10-20
What happened
STA released first results from Australia's National AI in Research Survey, run with MetaMelb at the University of Melbourne, which collected 1,034 complete responses across disciplines and career stages. Nearly three-quarters of respondents reported using AI (coding, literature search, specialised scientific applications and generative tools). More than half (54 per cent) considered their organisation's AI training inadequate, and institutional guidance was often described as unclear or poorly coordinated. Researchers generally accepted AI help drafting and editing grant applications but considered using AI to set funding scores or assess researcher or institutional performance unacceptable. Security and privacy concerns were the leading reported barrier to using AI-related computing; fewer than one in ten respondents quantified 12-month compute needs, but those estimates totalled about 53.5 million CPU-hours and 14.3 million GPU-hours. STA notes 2026 national allocation applications sought more than 2.2 billion hours, almost three times public capacity, and says commercial capacity should complement, not replace, public investment (NCRIS, NCI, Pawsey).
Why it matters
Gives research offices and funders a large, recent, cross-disciplinary baseline on researcher AI use, training gaps, acceptable-use boundaries in grant review and compute barriers. U.S. institutions designing AI training, research-computing allocation, and policies on AI in grant review or performance assessment can use it as a comparator, with jurisdictional limits.
Evidence and measured results
Opened and read the STA media release (20 Oct 2026) in full and the linked STA opening statement to the Joint Select Committee on AI (20 Oct 2026). Recorded the 1,034 responses, ~three-quarters using AI, 54% training inadequate (release), security/privacy as leading compute barrier, <10% quantifying compute, 53.5M CPU-hours and 14.3M GPU-hours, and the 2.2 billion requested hours. The opening statement adds that 80% of responses were from higher education, 70% reported moderate-to-extensive generative-AI experience, and the survey closed 25 September; it gives 45% (not 54%) for inadequate training. The linked initial-analysis PDF could not be retrieved (TLS error from the box).
Limitations and uncertainty
Initial analysis released by an advocacy peak body alongside parliamentary testimony; self-selected survey, no response rate or weighting reported in the release. The two STA documents differ on the inadequate-training share (54% in the release, 45% in the opening statement); the underlying PDF was not opened. Compute estimates come from fewer than 10% of respondents and STA says they are not a national forecast.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-23; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Researchers already use AI widely but lack adequate training, clear guidance and secure compute, creating integrity and data risks.
- Stakeholders
- Deputy vice-chancellor or vice president for research
- Research computing director
- Research integrity office
- Library and research-skills team
- Grants office
- Discovery questions
- Do you have local data on researcher AI use and training gaps?
- Where can researchers run AI on sensitive data today?
- Does policy address AI in grant scoring or performance review?
- Value hypothesis
- A local baseline and targeted training plus secure compute access could reduce unsafe workarounds and clarify acceptable use.
- Potential engagement
- A six-week researcher AI readiness baseline (survey plus interviews) with recommendations on training, guidance and secure compute.
- Unsupported claims
- AI raises research productivity by a measured amount
- Australian results transfer directly to U.S. campuses
- Compute estimates represent total demand
Pre-sales engineering
Role takeaway
- Fit
- Fits institutions planning secure research environments with GPU access and needing evidence on barriers.
- Architecture and integration
- Integrate a governed secure research environment with GPU partitions, identity controls and data-classification-aware access.
- Prerequisites
- Data-classification policy
- GPU capacity or burst arrangement
- Researcher intake and sizing process
- Constraints
- Self-reported, small compute sample
- Sensitive-data rules vary by jurisdiction and funder
- Initial analysis only; full report not yet reviewed
- Security
- Security and privacy are the top reported barrier; design for controlled data ingress/egress, logging and approved-model lists.
- Proof of value
- Measure researcher uptake of the secure environment over a term
- Compare sized requests with actual usage
- Survey whether training changes self-reported confidence and safe practice
Delivery
Role takeaway
- Work
- Run a local baseline survey, launch role-based AI training, publish acceptable-use guidance and open a secure GPU intake path.
- Dependencies
- Research office sponsorship
- Secure environment availability
- Training content and staff
- Ownership
- Research office (policy) with research computing (platform) and library (training).
- Skills and adoption
- Majority-scale training gap means programmes should cover tool evaluation, output verification and data protection, as STA describes.
- Governance checkpoints
- Research integrity review of acceptable-use guidance
- Data-protection sign-off for the secure environment
- Policy decision barring AI-determined funding or performance scores
- Acceptance criteria
- Proposed: agreed share of active researchers complete core AI training within a year
- Proposed: secure-environment onboarding time under an agreed target
- Proposed: guidance published and referenced in grant processes
- Risks
- Survey self-selection bias
- Training that lags tool change
- Conflicting figures across source documents
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Compute, secure data, storage and specialist workforce are framed as one planning unit; security and privacy are the main barrier to AI computing, pointing to secure research environments rather than raw capacity alone.
Governance
Who approves, reviews and stays accountable for outcomes?
Respondents draw a clear line: AI may assist drafting grant applications but should not determine funding scores or performance assessments. Institutions and funders can encode that boundary in policy.
Security and privacy
What data, permissions and controls need testing?
Security and privacy concerns are the top reported barrier to AI computing use; researchers want support to protect sensitive information.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
A majority-scale training gap and unclear guidance suggest institutions need structured AI skills programmes and expert support staff, not just tool access.
Procurement
What should contracts, pricing and exit terms secure?
STA positions commercial and hyperscaler capacity as a complement to sustained public research infrastructure, and expects firms operating locally to contribute to R&D capability.
Operating model
Which teams own the service once it runs?
Pairs research-computing planning with training, secure data services and research-integrity guidance under one roadmap (the 2026 National Research Infrastructure Roadmap).
Publication history
- 2026-10-22Research Universities · Issue 445 resources
Stable resource ID: sta-national-ai-in-research-survey-compute-training-20261020