Using AI responsibly and effectively: Building capability, confidence, and ethical use of AI in training.

Duration: 4m20s

Artificial Intelligence, or AI, is changing the way people work, write and learn.

The question for trainers is not whether to use AI. Rather, the question is how to use it effectively.

Used properly, AI can save time, support learning and help trainers to work smarter.

However, when it's not used well, it can lead to errors, poor judgement and serious integrity issues.

Remember, AI is a tool. It can help with developing ideas, building structure, crafting summaries, as well as drafting wording for resources and assessment questions.

But it shouldn't replace trainer judgement, hands-on skill and evidence of learner competence.

The value of AI depends on how well the trainer uses it, checks it and understands its limitations.

I like to think of it as the assistant, rather than the manager. Which means it needs supervision and feedback to progress successfully.

One of the best ways trainers can use AI is in preparation. 

It can help you to draft session outlines, discussion questions, learning resources and a host of other materials that support training.

But don't just copy and paste AI responses without checking them. AI can be wrong (and often is), as many of the platforms are designed to please the user, so they can confidently give you the answers that you want. As such, oversight is necessary.

AI responses can also miss context. And produce content that really doesn't fit your site or learner group.

Therefore, use AI to create a starting point, then apply your own expertise.

For trainers in mining and automotive, AI works best when it supports real work, not generic content.

This means you still need real tools, real tasks, real risks and real decisions in the training.

Of course, AI can help to shape the session. But the training still needs to connect to the job.

If the learning feels too generic, learners will disengage fast.

Learners also need guidance.

If AI use is allowed, be clear about what's acceptable and what must be declared.

For example, AI may help a learner brainstorm ideas, improve wording or summarise notes. But it shouldn't replace their own thinking, practical skills, or evidence of competence.

Make the rules clear early, and that'll help build clarity and confidence while protecting the integrity of the training.

If learners use AI, checking becomes even more important.

Don't assess the final document alone. Ask questions. Check understanding. Observe the task. Get the learner to explain their choices in their own words.

This helps you to confirm whether the work reflects real understanding.

In the end, a strong assessment looks beyond polished wording. It looks for depth in application.

Responsible AI use also means talking about ethics.

Trainers and learners need to consider things like accuracy, privacy and bias.

Never put sensitive workplace information into public tools without approval, and don't trust AI output just because it sounds legit.

The best way to build confidence is to model good use.

Show learners how to write a better prompt. Remember "rubbish in, is rubbish out". So, show them how to test their output and explain how they can make improvements.

That's how you build capability, rather than dependence.

AI can be useful, but we do need to exercise care.

Use it to support planning, improve learning and enhance the overall training and assessment experience. But always check the output.

Furthermore, set clear boundaries and keep real skill, real judgement and real performance at the centre of your practice.

That's how trainers can use AI responsibly and effectively. It's about getting yourself into good habits that ultimately rub off on your learners. Good luck with the challenge.

END.

Episode 13: Managing learner use of AI in assessment: Maintaining assessment integrity while enabling appropriate and transparent AI use.

Duration: 7m30s

AUSMASA Intro: Welcome to the Train-the-Trainer Podcast, proudly brought to you by AUSMASA, Empowering industry to develop essential workforce capabilities for today and tomorrow.

Howdy folks. I hope you're having a great day. My name's Marc Ratcliffe, and I'm your host again on the Train-the-trainer podcast. In this episode, we explore an emerging topic that's becoming more and more relevant across training and assessment: managing learner use of artificial intelligence, or AI, within assessment. Joining me in the discussion today is Kerri Sharpe. Hi Kerri!

Kerri: Hi Marc. This is a real hot-button topic, particularly with the regular, and I look forward to teasing out the major issues and actions with you today.  

Marc: For trainers and assessors in mining and automotive, this can feel like unfamiliar territory. Kerri, I know that AI tools are now easy to access and increasingly common in everyday work and study, but is it really impacting assessment in this environment?

Kerri: The short answer is yes. The question is not simply whether learners should use AI or not – the reality is they are. The more useful question is 'what kind of use is appropriate, and how do we manage it transparently?'

Marc: Okay. So, what do you see as the biggest issues?
 
Kerri: Let's begin with the core issue, Marc: assessment integrity. If a learner submits work that doesn't genuinely represent their own competence, the assessment result becomes unreliable. For example, blasting is a core activity in mining, but it's strictly regulated to protect people, property, and the environment. If we are testing learners' understanding of the regulations or related safety Acts, but it's really the AI giving the responses, the impact could be diabolical. 

Marc: So true, Kerri. That integrity piece matters in any field but is especially important in environments where training links directly to safety and compliance, or where there is a licensing outcome. 

Kerri: Correct. At the same time, completely ignoring AI isn't a realistic solution. Many learners are already using it to brainstorm, summarise, draft, translate, and clarify information. 

Marc: And that can really add benefit to the learning process.

Kerri: Yeah. In many workplaces, AI-assisted tools are becoming part of normal practice and including within training and development. They're being used to draft content, create workplace-specific images, and even assist with coaching activities.

Marc: So, rather than reacting with fear or blanket assumptions, trainers need a clear and practical approach?

Kerri: Yes. I like thinking about AI as an 'assistant' rather than the 'manager'. As such, the assistant should be provided with supervision and feedback rather than being let loose unchecked. 

Marc: And we can model that ethical practice too as trainers and assessors.

Kerri. Yeah. I think the first step is to set expectations. Learners should know what is allowed, what is not allowed, and what sorts of declarations they need to make if they are using AI.

Marc: Can you give an example, Kerri?

Kerri: Sure. For instance, AI might be used to help plan an answer or improve grammar or sentence structure, but not to generate final responses without disclosure.

Marc: Yeah, I think this clarity prevents confusion, and it also gives assessors something concrete to refer back to, should a student overstep. 

Kerri: Totally. But there are plenty of appropriate ways AI can be used to add value to the training and assessment experience. One I saw recently was where the trainer invited the learners to feed a summary of a common site incident into ChatGPT and ask it to suggest the steps to manage it. But the students' assessment was to review the AI's response and evaluate how effective it would be based on their company's policies and procedures.

Marc: I guess that would really help to check their depth of knowledge, rather than just cut and pasting answers—nice one.

Kerri: Thanks. 

Marc: Okay, so what's next on your hit list, Kerri?

Kerri: Well, the next step is to focus on evidence of actual competence. This often means relying less on take-home written responses alone and more on methods such as observation, verbal questioning, scenario discussions, and application in real or simulated work contexts. 

Marc: So, the more directly we see the learner perform, explain, and make decisions on the job, the less authenticity risk there is?

Kerri: Yes. Transparency is the key here. If learners are allowed to use AI in some way, ask them to explain how they used it. What tool did they use? For what purpose? What parts did they change or verify themselves? This encourages responsible use and also speaks to their digital literacy.

Marc: But it's also important to talk about the limitations of AI, right?

Kerri. Yeah, I agree. These tools can produce convincing but inaccurate information. They may generate generic answers that sound competent but do not reflect the local site procedures or current requirements. Learners need to understand that using AI without checking it can create risk, not just in assessment, but in the workplace.

Marc: So, this could be another teachable moment? 

Kerri. Absolutely. 

Marc: Okay, so how can we detect AI use when it's not declared?

Kerri: Consistent patterns in responses can be an indicator. 

Marc: What do you mean?

Kerri: Well, AI will often produce outputs that reuse the same phrasing, are overly generic or include an unusually long set of bullet points.

Marc: So, more fluff than substance?

Kerri: Something like that. Importantly, if a response seems inconsistent with the learner's usual level of performance, or disconnected from the task context, you may need to validate their knowledge another way – perhaps through verbal questioning or demonstration as part of a challenge test. 

Marc: That makes sense. 

Kerri: However, it's important that assessment decisions are evidence-based, rather than suspicion-based.

Marc: Yeah, and I think there is also a broader opportunity here for us as trainers and assessors. AI can actually become part of the learning conversation. We can discuss when it may be useful, when it may be risky, and how to use it ethically. 

Kerri: I think you're right, Marc. Those types of actions are becoming increasingly relevant as workplaces adopt more digital tools.

Marc: For sure. 

Kerri: But I think it is important to remember that we are not trying to create a cat-and-mouse game here. The aim is to protect the integrity of the assessment process while preparing learners to act responsibly in a changing environment.

Marc: And it's definitely a rapidly moving space, Kerri! Well, you have given us plenty of thought-provoking things to consider today. Thanks for being part of the Train-the-trainer podcast.

Kerri: You're welcome, Marc. If I can just close by inviting listeners to get out of their comfort zones and take a look at some of the AI tools out there. There are plenty of free trails on offer and you can start to familiarise yourself with what the learners are going to using.

Marc: That's good advice, Kerri. IN this way, they can pre-arm themselves with some of those common outputs they're likely to see.

If you're managing AI in assessment, keep it practical. Set clear rules. Design assessments that gather direct evidence. Require transparency where AI use is permitted. Check understanding through human interaction. And remind learners that competence cannot be outsourced.

Because at the end of the day, assessment is about confidence in performance. And in mining and automotive, that confidence needs to be real.

Thanks for your time.

AUSMASA Outro: This was just one resource in the AUSMASA Train-the-trainer suite of tools aimed at bridging the gap for trainers and assessors in the mining and automotive industries. Check out the other learning assets to take your training and assessment to the next level, including videos, scenarios, case studies, job aids, fact sheets and other podcast episodes. 

END.

Using AI responsibly and effectively: Building capability, confidence, and ethical use of AI in training.

Duration: 4m20s

Artificial Intelligence, or AI, is changing the way people work, write and learn.

The question for trainers is not whether to use AI. Rather, the question is how to use it effectively.

Used properly, AI can save time, support learning and help trainers to work smarter.

However, when it's not used well, it can lead to errors, poor judgement and serious integrity issues.

Remember, AI is a tool. It can help with developing ideas, building structure, crafting summaries, as well as drafting wording for resources and assessment questions.

But it shouldn't replace trainer judgement, hands-on skill and evidence of learner competence.

The value of AI depends on how well the trainer uses it, checks it and understands its limitations.

I like to think of it as the assistant, rather than the manager. Which means it needs supervision and feedback to progress successfully.

One of the best ways trainers can use AI is in preparation. 

It can help you to draft session outlines, discussion questions, learning resources and a host of other materials that support training.

But don't just copy and paste AI responses without checking them. AI can be wrong (and often is), as many of the platforms are designed to please the user, so they can confidently give you the answers that you want. As such, oversight is necessary.

AI responses can also miss context. And produce content that really doesn't fit your site or learner group.

Therefore, use AI to create a starting point, then apply your own expertise.

For trainers in mining and automotive, AI works best when it supports real work, not generic content.

This means you still need real tools, real tasks, real risks and real decisions in the training.

Of course, AI can help to shape the session. But the training still needs to connect to the job.

If the learning feels too generic, learners will disengage fast.

Learners also need guidance.

If AI use is allowed, be clear about what's acceptable and what must be declared.

For example, AI may help a learner brainstorm ideas, improve wording or summarise notes. But it shouldn't replace their own thinking, practical skills, or evidence of competence.

Make the rules clear early, and that'll help build clarity and confidence while protecting the integrity of the training.

If learners use AI, checking becomes even more important.

Don't assess the final document alone. Ask questions. Check understanding. Observe the task. Get the learner to explain their choices in their own words.

This helps you to confirm whether the work reflects real understanding.

In the end, a strong assessment looks beyond polished wording. It looks for depth in application.

Responsible AI use also means talking about ethics.

Trainers and learners need to consider things like accuracy, privacy and bias.

Never put sensitive workplace information into public tools without approval, and don't trust AI output just because it sounds legit.

The best way to build confidence is to model good use.

Show learners how to write a better prompt. Remember "rubbish in, is rubbish out". So, show them how to test their output and explain how they can make improvements.

That's how you build capability, rather than dependence.

AI can be useful, but we do need to exercise care.

Use it to support planning, improve learning and enhance the overall training and assessment experience. But always check the output.

Furthermore, set clear boundaries and keep real skill, real judgement and real performance at the centre of your practice.

That's how trainers can use AI responsibly and effectively. It's about getting yourself into good habits that ultimately rub off on your learners. Good luck with the challenge.

END.