Prompt Engineering Meets Training Package Design: What RTOs Must Know About AI Literacy in 2026

A comprehensive analysis of what AI literacy now means for Australian RTOs: the skills data putting prompt engineering among the nation's fastest-growing capabilities, the ICT Training Package redesign delivering new AI qualifications between now and 2028, the 2025 Standards obligations that AI-integrated delivery must satisfy, ASQA's draft principles for responsible AI, and the five actions providers should take before the regulator asks.

The Signal RTOs Cannot Ignore

LinkedIn's Skills on the Rise 2026 report placed prompt engineering, data analysis and insights, and AI model training and fine-tuning among Australia's fastest-growing professional skills, alongside stakeholder collaboration, client communication, cross-cultural capability and governance. For the vocational education and training sector, this is not a passing headline. It is a signal (regulatory, pedagogical and strategic) that RTOs can no longer treat artificial intelligence as a niche ICT elective or an afterthought in training and assessment design.

The timing matters. The Standards for RTOs 2025 have been in force since 1 July 2025, and ASQA's regulatory framework has shifted decisively toward outcomes, self-assurance and continuous improvement. The Future Skills Organisation, the Jobs and Skills Council for finance, technology and business, is redesigning the ICT Training Package to address urgent skills gaps in artificial intelligence and cyber security. And in May 2026, ASQA used its Sector Webinar to share draft principles for the responsible use of AI in VET, the clearest indication yet that AI has moved from an innovation conversation to a standards conversation.

The RTOs best positioned for 2026 and beyond are those proactively embedding AI literacy into their training products, assessment strategies, governance frameworks and trainer professional development, rather than waiting for regulatory direction after the fact. The framework exists. The content is being built. The accountability has arrived.

This article provides a comprehensive analysis of what AI literacy means for RTOs in 2026, structured around five areas: the national policy context, training package reform, unit of competency and assessment design, ASQA expectations under the 2025 Standards, and the practical actions providers should take now.

1. The National Policy Context: Why AI Literacy Matters Now

1.1 The National Skills Agreement and the National Skills Plan 2025 to 2026 Update

The policy architecture is already in place. The National Skills Agreement, the five-year joint agreement between the Commonwealth and the states and territories that commenced on 1 January 2024, names "ensuring Australia's digital and technological capability" among its national priorities, alongside reforms to improve the regulation of VET qualifications and quality. The National Skills Plan 2025 to 2026 Update, published by the Department of Employment and Workplace Relations, details progress against those priorities and sets new focus areas for 2026, with digital capability, data systems and cyber capability woven through workforce planning and funding directions.

The implication for providers is structural, not rhetorical. Digital and AI literacy are no longer confined to information and communications technology qualifications. They are cross-sectoral competencies that touch business, health, community services, construction, manufacturing and every other industry area delivered through the national training system. Australia's skills system is expected to produce graduates who can work effectively with AI tools, understand their limitations, and apply human judgment in AI-augmented workplaces. RTOs that fail to reflect this in their training products risk delivering qualifications misaligned with industry expectations, government priorities and, increasingly, regulatory scrutiny.

1.2 The Digital Education Council AI Literacy Framework: A Useful Lens, Translated for VET

The Digital Education Council's AI Literacy Framework, released in 2025, offers a structured, human-centred model of what AI literacy actually means. It defines five dimensions, each with three competency levels running from awareness through application to higher-order capability. The framework was developed primarily for higher education, but it translates directly into vocational contexts, and the translation exercise is precisely the kind of cohort-level thinking the 2025 Standards reward.

The framework is most useful for answering a question every RTO should now be asking at the training product level: what does "AI literate" mean for this learner, at this qualification level, in this industry? A Certificate III graduate in allied health assistance needs different AI capabilities from a Diploma graduate in project management or a Certificate IV learner in training and assessment. The common thread is that all graduates should understand what AI can and cannot do, how to use AI tools ethically and effectively in their workplace, and how to critically evaluate AI-generated outputs.

2. Training Package Reform: The AI Content Is Being Built Now

2.1 The ICT Training Package Redesign

The Future Skills Organisation is redesigning the ICT Training Package in response to its needs and gaps analysis, Towards Effective ICT Training, which found that current training products no longer meet industry needs, lacking both clear entry-level pathways and a workable balance between generalist digital skills and specialist ICT streams. The scale of what is being reworked deserves attention: the ICT Training Package comprises 11 qualifications, 108 skill sets and 650 units of competency, and is used by around 1,500 RTOs across Australia.

The redesign will deliver a new suite of qualifications, units of competency and skill sets that clearly separate generalist digital skills from specialist ICT streams, with clear job role pathways and stackable credential options. Specialist artificial intelligence and specialist cyber security training products are already moving through drafting and stakeholder consultation during 2026. Critically for the wider sector, the new AI-aligned products are being designed for flexibility and contextualisation across diverse delivery environments, which opens the door for RTOs delivering non-ICT qualifications to import AI-focused units and skill sets into business, health, construction and other training products.

2.2 The Qualification-First Model and AI-Infused Outcomes

National qualification reform is shifting VET design away from a unit-by-unit approach toward purpose-led, qualification-level design, with stronger emphasis on holistic outcomes and clearer pathways across VET and higher education. For AI and digital capability, the consequence is that AI literacy should be designed into the overall qualification outcome (defining what a competent graduate can do with AI in the workplace) and then expressed through clustered units, integrated assessments and workplace learning, rather than bolted onto a single elective.

For RTOs, the practical discipline is a single question asked at every training product review: what does a graduate of this qualification need to be able to do with AI tools in their industry? The answer will vary by sector. The question should not.

2.3 Cross-Package Application: AI Beyond ICT

One of the most significant implications of the redesign is cross-package use. A Certificate IV in Business could include an AI skill set covering the use of AI tools for data analysis, document generation and decision support. A Diploma of Nursing could incorporate capability in AI-assisted clinical documentation and the ethical boundaries of AI in patient care. A construction qualification could address AI-supported safety data analysis and reporting.

RTOs should anticipate this cross-package use in their training and assessment design now, rather than treating AI as the exclusive property of ICT qualifications. The industry signal is consistent across sectors: employers increasingly expect graduates to arrive with practical AI capability, and providers that respond early will hold an advantage in both learner attraction and employer satisfaction, the very measure on which national data shows the sector can least afford further decline.

3. Units of Competency and Assessment Design: Where Prompt Engineering Fits

3.1 From "Using Software" to "Working with AI Systems"

Even before the new AI-specific products formally land in training packages, RTOs can and should interpret existing digital, problem-solving and information skills through an AI lens. The Australian Digital Capability Framework provides the structure. It organises digital capability into five focus areas (information and data literacy; communication and collaboration; digital content creation; protection and safety; and technical proficiency and problem solving), assessed from foundation through to specialised levels, and it is increasingly used by RTOs preparing learners and systems for the 2025 Standards environment.

Within this framework, prompt engineering is best positioned as the practical, user-side expression of AI literacy: how learners, trainers and employers instruct, question, and critically evaluate AI tools in real vocational tasks. This is not about turning every learner into a software developer. It is about ensuring graduates can use AI tools effectively and responsibly in their specific workplace context (drafting clinical notes, generating project reports, analysing safety data, supporting customer communications) and can recognise when the tool is wrong.

3.2 Embedding Prompt Engineering Across the Evidence Dimensions

When designing or updating assessment tools and supporting resources, RTOs can embed prompt engineering and AI literacy across the familiar dimensions of a unit of competency:

3.3 Assessment Integrity in an AI-Enabled Environment

The integration of AI into assessment raises integrity questions that RTOs must address proactively, because the regulator already is. ASQA's Corporate Plan 2025 to 2026 emphasises protecting the integrity of the VET system, including action against non-genuine assessment evidence and non-authentic student work. And ASQA's regulatory casework is now surfacing AI-specific failures: cohorts submitting near-identical AI-generated responses, providers unprepared for student misuse of AI tools, and compliance documentation drafted by AI that bears no relationship to actual practice.

The practical implications are clear, and they map directly onto the 2025 Standards. Every provider's documented training and assessment approach must explicitly state whether and how learners may use AI in assessment, and the position should be granular: AI may be permitted for research and drafting with full attribution in some tasks, restricted to specific supervised activities in others, and entirely prohibited in high-risk competency demonstrations such as clinical skills or safety-critical operations. Assessment tools should include conditions of assessment that specifically address AI. Pre-use review of assessment tools under Outcome Standard 1.3, assessment conduct under Outcome Standard 1.4, validation under Outcome Standard 1.5 and systematic monitoring under Standard 4.4 must all consider AI-related risks, including the similarity of AI-generated responses and the challenge of authenticating learner work.

The anchor remains the rules of evidence under Standard 1.4: every assessment judgement must be justified against validity, sufficiency, authenticity and currency. If an assessor cannot distinguish between a learner's genuine competency demonstration and an AI-generated response, the assessment tool itself needs redesign. This is not about banning AI. It is about designing assessment conditions under which the evidence collected can still satisfy the rules of evidence that have always governed competency-based assessment.

4. ASQA Expectations: The 2025 Standards and the Regulator's 2026 Position on AI

4.1 Quality Area 1: Training and Assessment with Digital and AI Capability

From 1 July 2025, all ASQA-regulated RTOs must comply with the Standards for RTOs 2025, which shift emphasis from prescriptive compliance to high-quality outcomes, self-assurance and continuous improvement. Quality Area 1 requires engaging, well-structured, industry-relevant training and assessment that enables accurate judgments of competency, and ASQA's practice guides clarify expectations for contemporary learning methods, industry currency and the integration of technology where it supports outcomes.

For AI-integrated delivery, this translates into three core requirements. First, AI tools must deepen learning through simulations, data analysis, scenario building and practice activities, rather than shortcut the development of competence. Second, trainers and assessors must hold both vocational competency and current AI or digital skills where AI is part of delivery or assessment, evidenced through recent, relevant industry engagement and professional development; this sits squarely within Outcome Standards 3.2 and 3.3 and the Credential Policy, and ASQA has been explicit that industry currency, not the TAE credential, is where the sector's workforce compliance problems now concentrate. Third, documentation must show how AI use supports, rather than replaces, the learner's performance of the actual elements and performance criteria specified in the unit of competency.

It is worth stating the regulatory context plainly. ASQA's first eleven months of performance assessment under the 2025 Standards identified assessment quality (Standard 1.4), training quality (Standard 1.1), continuous improvement (Standard 4.4) and trainer and assessor industry currency (Standard 3.3) as the four most common areas of non-compliance. AI-integrated delivery touches all four. A provider that introduces AI into training and assessment without governing it is not innovating around the sector's biggest compliance risks. It is compounding them.

4.2 LLN, Digital Literacy and AI Readiness at Enrolment

The 2025 Standards expect RTOs to assess language, literacy, numeracy and digital capability before enrolment and to provide targeted support, so that learners are not enrolled into programs in which they cannot reasonably succeed. Commercially available LLN and digital literacy assessment tools already align with the 2025 Standards and the Australian Digital Capability Framework levels.

In AI-integrated delivery environments, the pre-training review should extend to AI-related digital readiness where relevant: basic search literacy, the ability to follow procedural instructions in digital environments, and awareness of privacy and security prompts. The results should inform induction content (such as an AI literacy orientation module for learners with low digital confidence) and reasonable adjustments that support learners to engage with AI-enhanced training without being disadvantaged. Equity is not a side issue here: ASQA's own draft AI principles make student equity, accessibility and wellbeing an explicit expectation of AI-enabled delivery.

4.3 Governance, Self-Assurance and ASQA's Draft AI Principles

ASQA's Corporate Plan 2025 to 2026 highlights risk-based regulation, data analytics and sector capability uplift, with self-assurance (not compliance assembled for audit day) as the operating expectation. Providers with strong, evidence-based self-assurance and quality outcomes may see reduced regulatory burden; those with poor practices will face more intensive scrutiny, and from 1 July 2026, that scrutiny carries fixed, published fees.

At its 2026 Sector Webinar, ASQA introduced draft principles for the responsible use of AI in the VET sector. The principles are guidance rather than new regulatory requirements at this stage (they tie into existing obligations under the 2025 Standards), but they describe exactly how the regulator expects providers to think.

For AI governance, this means AI policies must be part of the RTO's governance framework: approved by the governing body or leadership team, aligned to privacy, cyber security and academic integrity obligations, and regularly reviewed. Risk registers and internal audit plans should explicitly include AI-related risks: dependence on vendor tools, data leakage, undisclosed AI-assisted assessment, and staff skills gaps. Quality indicators and self-assurance evidence, including moderation records, student feedback and completion and outcome data, should be analysed for negative patterns linked to AI-mediated learning or assessment. And one warning from ASQA's own casework should be laminated and placed on every compliance manager's desk: the regulator has received annual declarations with the AI prompts still visible in the text. AI can assist with drafting. It cannot create compliance where the underlying system does not exist.

5. Five Actions RTOs Must Take in 2026

Action 1: Redesign training and assessment documentation and learning resources with AI literacy in mind

For each AI-relevant qualification or unit on scope, the provider's documented training and assessment approach should explicitly state how AI will be used in training, how it will be controlled in assessment, and how this aligns with unit and qualification outcomes. AI-related outcomes, including prompt engineering skills, should be mapped to performance criteria, knowledge evidence and foundation skills, with the mapping captured in validation records. Learning materials should teach safe, ethical and effective AI use: prompt design, output verification, documentation of AI use, and respect for intellectual property and privacy. The 2025 Standards no longer prescribe a training and assessment strategy template, but the obligation to demonstrate a coherent, documented, evidence-based delivery model is fully intact. AI use is now part of what that model must explain.

Action 2: Build trainer and assessor capability in AI

Under Outcome Standards 3.2 and 3.3 and the Credential Policy, trainers and assessors must hold required credentials and current industry skills and knowledge for the products they deliver. Where AI tools are integrated into delivery, the RTO must be able to evidence trainer currency in those tools and in related industry practice through professional development records, industry engagement and workplace projects. For prompt engineering specifically, professional development should cover how different AI models respond to instructions, context, examples and constraints; how to design prompts that reflect unit performance requirements and workplace contexts; and how to use AI outputs critically rather than accepting them at face value. Generic statements of digital confidence will not satisfy a trainer matrix. Mapped, current, product-specific evidence will.

Action 3: Update student-facing policies and academic integrity frameworks

Implement a clear "Use of AI in Student Assessments" policy that defines permitted, restricted and prohibited AI use, disclosure expectations and consequences for misuse. Communicate it at enrolment and reinforce it throughout training. Align it with privacy, cybersecurity and data-handling obligations, and update existing academic integrity and plagiarism policies to address AI-generated content as a distinct category alongside traditional plagiarism. A policy that students have never seen is not a control. It is a liability with a version number.

Action 4: Extend pre-training review to include AI and digital readiness

Where programs incorporate AI tools in delivery, extend LLN and digital capability assessment at enrolment to capture AI-related readiness: familiarity with digital search tools, ability to follow structured digital instructions, awareness of data privacy principles, and comfort with technology-mediated learning. Use the results to inform learner support planning, induction and bridging content, and reasonable adjustments, so that AI-enhanced delivery lifts the cohort rather than filtering it.

Action 5: Prepare for ASQA's risk-based scrutiny with evidence that already exists

Given ASQA's risk-based approach, its four published areas of concern and its growing interest in technology-enabled delivery, audit-ready evidence must exist as part of normal operations, not be assembled after a notification arrives. That evidence set includes training and assessment documentation, mapping and learning resources showing how AI is integrated without compromising unit requirements; AI policies and staff professional development records; pre-use review and validation records that address AI risks; AI-related complaints or incidents and their resolution; and proof that learners' digital and AI readiness was assessed at entry with appropriate support provided. The governance test is simple. Can the RTO show how AI use is controlled today, on request? If not, that is the gap to close first.

6. The Broader Picture: AI Skills as a Competitive Advantage

The Skills on the Rise 2026 data carries a pattern that should reassure rather than unsettle the VET sector. Australia's fastest-growing skills are not exclusively technical. Communication, stakeholder collaboration, cross-cultural capability, governance and ethics feature prominently alongside prompt engineering and model training. So, notably, does a cluster the report groups under training, coaching and people development, including mentoring, onboarding design and learning program design. The labour market is not only demanding AI skills. It is demanding the people who can teach them. The VET workforce itself appears in the growth data.

This convergence of technical and human capability reinforces what competency-based training has always been designed to deliver: graduates who can apply knowledge and skill in workplace contexts, make sound judgements, communicate effectively and adapt to changing conditions. As AI reduces the cost of generating technical output, the premium on human judgement, relationships and accountability increases. AI does not replace the educational mission of vocational education. It amplifies it. The RTO that produces a graduate who can both use AI tools effectively and exercise the critical thinking, ethical awareness and communication skills to use them responsibly will deliver the most value to employers, the stakeholder group whose satisfaction with VET the sector most urgently needs to rebuild.

Conclusion: The Framework Exists, the Content Is Coming, the Accountability Has Arrived

The message for RTOs in 2026 is unambiguous. AI literacy, including practical skills such as prompt engineering, is no longer confined to specialist ICT qualifications. It is a cross-sectoral capability expectation that must be reflected in how RTOs design training products, structure assessments, govern their operations, support learners and develop their workforce.

Every foundation is already in place. The Standards for RTOs 2025 provide the framework, and eleven months of regulatory evidence show exactly where it bites. The ICT Training Package redesign provides the content, with specialist AI products in consultation now and delivery running through to 2028. ASQA's draft AI principles and risk-based, fixed-fee regulatory model provide accountability. The national skills data provides the demand signal.

What remains is for RTOs to act: deliberately, strategically, and with the self-assurance mindset the new regulatory environment demands. The signal could not be clearer. What the regulator, the Jobs and Skills Councils and the labour market are now waiting to see is the response.

References and Further Reading

ASQA (2025). Standards for RTOs 2025. Australian Skills Quality Authority. https://www.asqa.gov.au/rtos/2025-standards-rtos

ASQA (2025). Practice Guides for the Standards for RTOs 2025. https://www.asqa.gov.au/rtos/2025-standards-rtos/practice-guides

ASQA (2025). Corporate Plan 2025 to 2026. Australian Skills Quality Authority. https://www.asqa.gov.au

ASQA (2025 to 2026). Artificial Intelligence (AI) Transparency Statement. https://www.asqa.gov.au/about-us/governance-and-accountability/artificial-intelligence-ai-transparency-statement

ASQA (2026). 2026 ASQA Sector Webinar, including draft principles for the responsible use of AI in VET. https://www.asqa.gov.au/newsroom/latest-news/2026-asqa-sector-webinar

Department of Employment and Workplace Relations (2026). National Skills Plan 2025 to 2026 Update. https://www.dewr.gov.au/national-skills-agreement/resources/national-skills-plan-2025-26-update

Department of Employment and Workplace Relations. National Skills Agreement. https://www.dewr.gov.au/national-skills-agreement

Digital Education Council (2025). AI Literacy Framework. https://www.digitaleducationcouncil.com

Future Skills Organisation (2025 to 2026). ICT Training Package Update: Towards Effective ICT Training: A Needs and Gaps Analysis of the ICT Training Package. https://www.futureskillsorganisation.com.au/ict-training-package-update/

LinkedIn (2026). Skills on the Rise 2026: The Fastest-Growing Skills. https://news.linkedin.com/2026/Skills-on-the-rise-2026

National Vocational Education and Training Regulator (Outcome Standards for NVR Registered Training Organisations) Instrument 2025. Federal Register of Legislation.

Credential Policy, Revised Standards for RTOs. National Training Register. https://content.training.gov.au/sites/default/files/2025-03/Credential%20Policy.pdf