October 7, 2025 • 33 mins
The vocational education landscape is experiencing a seismic shift as artificial intelligence transforms everything from assessment design to compliance management. This thought-provoking conversation with AI expert Lauren explores the practical implications of this revolution for RTOs.
We dive deep into how AI is already reshaping assessment validation processes, with Lauren sharing how her organisation has redesigned hundreds of assessments to incorporate live video components. This critical adaptation allows trainers to validate a student's written submissions against their verbal capabilities—helping to identify potential AI-generated work when written language significantly outpaces verbal expression.
The discussion takes a cautionary turn as we explore the vital questions RTOs should ask before adopting any AI solution. Where is your data being stored? What security protocols protect it? Most importantly, do the people behind AI compliance tools actually understand RTO operations and regulatory requirements? Lauren emphasizes that no matter how impressive the technology, the human expertise behind it remains essential.
For RTOs yet to begin their AI journey, we offer practical starting points based on your scope of registration. Those in business, finance or IT sectors face the stark reality that students are already using AI extensively. Rather than futile resistance, we suggest strategies for incorporating AI authentically into assessments while maintaining academic integrity.
Perhaps most concerning is Lauren's prediction about the future of compliance challenges. While ASQA has recently cancelled thousands of qualifications due to insufficient evidence, tomorrow's challenge won't be lack of evidence—it will be determining authenticity as AI tools become sophisticated enough to generate convinc
Join host Angela Connell-Richards as she opens each episode with a burst of insight and inspiration. Discover why compliance is your launchpad to success, not a limitation.
Connect with fellow RTO professionals in our free Facebook groups: the RTO Community and RTO Job Board. Visit rtosuperhero.au/groups to join today.
Wrap up with gratitude and guidance. Subscribe, leave a review, and join our community as we continue supporting your compliance journey in vocational education.
Thank you for tuning in to the RTO Superhero Podcast!
We’re excited to have you join us as we focus on the Revised Standards for RTOs in 2025. Together, we’ll explore key changes, compliance strategies, and actionable insights to help your RTO thrive under the new standards.
Stay connected with the RTO Community:
📌 Don’t forget to:
âś” Subscribe to the RTO Superhero Podcast so you never miss an episode!
✔ Share this episode with your RTO network—compliance is a team effort!
🎙 Listen now and get ahead of the compliance changes before it’s too late!
📢 Want even more compliance insights? Subscribe to our EduStream YouTube Channel for our FAQ series on the New Standards for RTOs 2025! 🎥
đź”— Subscribe now: EduStream by Vivacity Coaching
✉️ Email us at hello@vivacity.com.au
đź“ž Call us on 1300 729 455
🖥️ Visit us at vivacity.com.au
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:03):
with me, Angela Connell-Richards, and once again
with Lauren, who's joining mein a great episode, one of our
favourite topics, which is AI.
Welcome Lauren, how are you?
Speaker 2 (00:16):
Good, good.
Thank you so much for having meagain.
Speaker 1 (00:19):
Yes, it's awesome.
So we were chatting offlineabout this and about AI and
where the focus is going in thefuture, in particular, when it
comes to how ASQA are going toaudit us and the different
direction that businesses aregoing in due to AI and that type
(00:43):
of chat, gpt and technology, inparticular, resources and
things like that.
So tell me what drew you to AIfrom the beginning and, in
particular, how it applies inthe LTO sector.
Speaker 2 (00:58):
Yeah, so I did my first talk on AI in 2020, I
think it was 2023, velk was thefirst like public speech that I
did about that.
Obviously, we've been using AIlike GPT, claw, bard, grok,
things like that for that sameperiod of time, watching it,
(01:21):
watching its kind of itsevolution.
And then I was at a conferencea couple of weeks ago, the VETQI
conference and certainly therewas a huge.
You know, there was a lot ofquestions about it, topics about
it, and I think when we haveVELG this year, the VELG
conference I think we're goingto see a lot of speakers talking
(01:42):
about that.
I think there's going to be alot of questions and you know
we've got these new standards inwhich have a really strong
focus on governance, and when Istart to look at how RTOs are
engaging with AI, how trainersand assessors are engaging with
AI, how students are engagingwith AI, I just think it's one
(02:06):
of those spaces that you know issubstantially going to change
how we work.
You know I homeschool my childand you know I had Zach read a
book last week and I said to himokay, I said I want you to sit
down and chat GPT, I want you tokind of you know, create a
structure for a book report.
(02:28):
I want you to cover X, y and Zand then you know you'll need to
work with it to refine thereport.
This is a kid that's got youknow dyslexia and you know hates
writing.
But you know that's a reallyeasy way for him to be able to
interact so that he can create aproper report and effectively
that's how he would create anysort of report or document in
(02:49):
the workplace, right, yeah?
So, you know, we're now, then,starting to look at how
certainly we're having to lookat how we adapt our assessment
tools.
Um, so, in Annawire we build alot of our assessments in
Accelerate for our clients andwe've just redesigned I think
over the last two months we'vehad to rebuild something like
(03:11):
700 or 800 assessments in ourclient systems because we're now
starting to incorporate a livevideo component.
So when we do our validations,we can validate what the student
is typing, like their writtenlanguage, in comparison with
what they can produce verballyand how they can talk to a topic
(03:34):
verbally.
When you can see a student whois writing very academically but
(03:55):
when they speak, their orallanguage would demonstrate an
AQF level two to three levelsbelow what they're talking to
right, same questions, but thewritten version and the oral
versions just aren't lining up.
And that's for us, that'ssomething that we're having to
start to do, because really,when it comes down to validation
, it boils down to okay, but ifyou're going to walk the walk,
(04:19):
you've got to talk the talk aswell, and I think this is
something we're going to startto see Now.
That easily then starts tobring in things like well, we
had to adapt our assessmentinstructions to go.
A live camera is going to comeon and it's going to record you
and please make sure there'snobody in the background and
please make sure you're dressedand professional in your
(04:41):
presentation, and you know.
So we then start to think about, from a privacy perspective,
what can and can't we do.
I know when I was at theconference I was talking to
Accelerate and they had a clientcome to them and go look, can
we create a software where wecan record in the childcare
centres and it automaticallyblows out all of the children's
(05:03):
faces and identities, carecentres and it automatically
blows out all of the children'sfaces and identities?
You know.
So AI is impacting all of thisright and there is so many
different ways in which itaffects people.
You know, as you know, mylittle AI Fireflies follows me
around to all of my meetings andin most of them she's allowed
to participate.
(05:23):
But certainly in any of mygovernment meetings I just turn
around to him, go look, I've gotno choice.
She follows me everywhere.
Just kick her out, you know.
But it's, it's a fantastic toolbecause you know now, when I
used to struggle to todemonstrate all the quality
activities I was doing withclients.
Now, every meeting is recorded,meeting minutes are produced
(05:46):
and then my amazing operationsmanager, laura, takes every
single one of those meetingnotes and copies it into the
client's quality register.
And so whenever we're going toask for and saying what are the
quality activities you're doing,we've got know notes upon notes
upon notes upon notes showingthat every couple of weeks we're
(06:08):
going, we're make, we'retalking about an improvement
we're making or we'reimplementing a plan, and it's
all there and it's so easy, youknow.
So I think what ai could do isabsolutely brilliant in our
sector.
Um, it certainly makes iteasier for me to make far more
innovative and enjoyable andengaging assessments and content
(06:29):
.
But there's a whole other sideof it as well, and you know we
were talking about before.
You know like you've got an AIproduct.
There's three or four of theseAI compliance hubs out there
from, you know, little teenyboppers no offense to the teeny
boppers out there, but you knowlike the photos look like photos
(06:49):
of my 18 year old.
You know who's clearly thislittle group of boys that have
got this great AI knowledge, butnow they're saying that they
can help you run your wholebusiness, but they've never
actually worked in vet beforeyou know.
So there's a lot of stuff outthere that RTOs are now going to
get kind of oh well, you coulddo this and you can do this.
(07:11):
And it's that filtering process.
Again, you know our kidsgrowing up nowadays they don't
have to.
It's not about whether or notthey've got access to the
information, it's filtering theinformation.
What's the quality of it andwhere does it come from?
And I think RTOs have to bejust as savvy.
What is the tool that I'mlooking at?
(07:32):
What is it built on?
Who is it coming from?
Right, and, and is thisactually something that's going
to save my business money andgive me a return on investment?
Or, you know, am I just buyingsomething that I think is going
to tick a box when really it'snot actually going to make any
(07:54):
sort of measurable impact on mybusiness?
Speaker 1 (07:55):
Yes, so what do you think are the biggest risks for
RTOs when it comes to adoptingAI?
Speaker 2 (08:03):
So I was having a really interesting conversation
with Vanessa, from Prickly toSweet, who's like she's been in
AI for what?
Eight or nine years now, andthen I think Kerry Buttery as
well, is another one that I'vehad a couple of chats with, and
she's very much operating inthat AI space too.
Where your data goes issomething that I've taken from
(08:24):
both of them.
Like that's definitelysomething what data you're being
asked to put into the AI andwhat data you're getting out of
it, and how it then secures it.
Is it being held in an offshoreserver or an Australian server?
Are there any?
That's the biggest question.
Speaker 1 (08:38):
I get asked.
Speaker 2 (08:40):
Yeah.
Speaker 1 (08:41):
When I refer people to Accelerate.
Speaker 2 (08:44):
I know they've got all all like.
They've got 27 different.
You know security controls intheir system, um, but you know
like and I mean I, you know thegovernment's using microsoft, so
microsoft and microsoftco-pilot and stuff like that
obviously their security andthings like that is approved,
but there's a lot of others that, like, trying to get that
information from them isproblematic, you know.
(09:07):
So I think that's one thingRTOs need to start to learn to
navigate and negotiate.
And then the Geico principle ofyou know people who have built
AI systems that don't have RTOknowledge, that can't
(09:29):
demonstrate that they've got RTOknowledge.
And you know not justnecessarily even RTO knowledge.
If you're looking at somethingto run your whole business and
your business runs six fundingcontracts, you know, okay, I can
write a policy set for your RTOstandards, but there is not a
single funding contract thatdoesn't require you to adapt
(09:50):
your policies for that fundingcontract.
You know, in terms of finance,in terms of how you capture your
data, in terms of yourenrolment process, in terms of
additional, you know things thatyou need to capture from the
students or information you needto provide to the students and
things like that.
So I just think that peopleneed to learn how to have really
(10:12):
good conversations with anyprovider that they're going to
look to use any sort of resourcefrom, and I would be super, I
would be very reluctant.
Um, and I've seen this, I'veseen this advertised on a few of
these RTO, um assessmentdevelopment things where they go
use this tool and you'll, wecan build, you'll, it'll, it'll
(10:35):
build your resources in fiveminutes.
It'll build a compliantcontextualized assessment in
five minutes.
And I'm just going to turnaround to people and say it's
bullshit.
Yeah, I have not seen a systemthat can write an assessment
tool in five minutes that'scompliant.
Most of them are built off ofgpt, which is built off of the
internet, and the internet isfull of non-compliant assessment
(10:57):
tools.
So you know, if it sounds toogood to be true, it's probably
too good to be true.
Speaker 1 (11:05):
So with ComplyHub so you've touched on that the
system that we've built, we'veactually integrated it with TGA
so it will actually bring in theinformation for the training
products that are on your scopeof registration.
So it does make a bigdifference.
But you know, even just youknow, having the unit of
(11:25):
competency and downloading thatfrom TGA and putting that into
chat GPT, you're not going toget a good assessment tool
because it's just going tofollow the legislation or the
requirements of the trainingproduct.
It's not going to go intoreal-world scenarios, it's going
to create an assessment againstPC 1.1 and one against PC 1.2.
Speaker 2 (11:46):
And you know what you might actually get like within
the next year.
You'll probably get maybe acompliant-ish tool, but it's not
going to be usable.
The assessors are going to hateit.
It's going to be overly, it'sgoing to probably be far more
academic than it should be forthe AQF level.
Like there's so many nuances,we still need humans Still need
(12:07):
humans.
Yes, thank you.
Yes, it's a great tool.
It's an great tool.
It's an amazing tool and I useit every day and my business
would be a lot more cumbersomeand more expensive without it.
I love it.
I think it's amazing what it'sgoing to do, but there has to be
that human component behind it,and anyone purchasing a service
(12:30):
that uses AI needs to talk tothe people that sit behind it.
So, like, if you're going to goand use Vanessa's AI validator,
go talk to Vanessa and see theknowledge that she has.
If they're going to use your AIcompliance hub, go and talk to
you and your team and see thecompliance that sits behind it.
I know that.
You know the learning resourcesgroup has got a couple of
(12:51):
different AI tools that they'rebringing out.
You know they've been in thisindustry and they've actually
built good, good assessmenttools and this is coming from
somebody who builds assessmenttools.
But, like we've always had a lotof respect for the learning
resources group because you knowthey don't sell these very
cheap assessments, you know thatthe contract requires you to.
(13:12):
You know, take seven stepsbefore you get any sort of
compliance guarantee out of themand things like that you know
they genuinely mean well, theygenuinely want to provide a
really good quality product, andyou know you need to talk to
the people that sit behind itand ask those questions, and the
(13:34):
governance standards are veryclear.
Do your due diligence.
Know the people that you'reworking with.
Know the companies, know thepeople that sit behind the
companies that you want toassociate yourself with.
Speaker 1 (13:44):
Yeah, yeah definitely , and I think one of the big
things around using AI withinyour like what I see is I need
it as part of what we do now asa service.
However, you still need to havea good understanding of the
requirements of the legislation,the standards, the different
(14:07):
other different requirementsthat you need to have in place
for a range of differentlegislation, and I think someone
trying to develop a compliancesystem who doesn't actually have
any idea about compliance isnot going to be able to build a
good system.
Like I'm now able to build asystem without a software
(14:30):
developer, and it's fantasticbecause I have built stuff with
software developers before.
It was always trying to teachthem what you do and how to do
it, whereas now I don't have todo that.
I'm doing it straight,one-on-one with the AI and I'm
teaching it and it remembers,which is great, and we're
building on that over and overand over again.
(14:52):
So it's, yeah, so much betterthan trying to work with a
software developer who's reallytech savvy but just really
doesn't understand the topicthat you're talking about, and I
totally agree with you therethat you really need to have a
good understanding of who you'reworking with and whoever's
building these systems?
Do they understand what are therequirements when it comes?
(15:16):
Because the last thing you wantis to turn up in an audit and
your auditor picks up somemistake that your AI, that AI
has made, put it at risk,particularly if you're getting
assessment tools built andputting your RTO at risk of
(15:36):
certificates being cancelled dueto not collecting sufficient
evidence.
Yeah, yeah, definitely so I'mgoing to move on to another
question.
For an RTO that hasn't startedyet with any AI, what would you
recommend would be the firstplace they should start?
Speaker 2 (15:58):
I would say it would depend on the scope of the RTO.
So I think for those that areoperating in more of the like
soft services, like business,financial IT, anything of that
sort of space, the first thingthat you need to know is that
your students are 100% using AInow, and if you're not, I can
almost guarantee that yourassessments are being
(16:21):
fraudulently submitted andyou're completely unaware of it.
So my first step in an RTO likethat would be to get the
assessors working with AI veryregularly, even if it's just
helping them to create theirsession plans.
Speaker 1 (16:36):
Contextualizing.
Speaker 2 (16:38):
Yep, contextualizing certain things, maybe you know,
getting three or fourassessments together and working
with AI to do the validation orsomething like that, just so
that they can start tounderstand the way that AI
writes the formatting that itprefers, the spelling that it
prefers, the spelling that itprefers, because that's going to
make it much easier for them tostart to pick up on that with
(16:59):
their students and also get themto start to think about how
they can update theirassessments to incorporate more
of AI.
Like I know, when we're writingpolicies now we're writing
assessments, now we'll dosomething like okay, you need to
write a policy, you need todevelop and implement a policy
on X, right?
What you're going to do isyou're going to work with AI to
(17:20):
develop a template all right anddevelop the draft, and then
you're going to go back andforth until you refine it and
get it to exactly what it is.
You're then going to submit acopy of the chat with your AI
and the finalized document,right?
That is exactly aligned withwhat we would expect in the
workplace.
(17:40):
We're not putting limits onstudents about working with AI,
but we are getting the AItranscript with the students so
that we can see what thestudent's actual knowledge is
right.
Yeah, and teaching trainers howto like, tweak and adapt
assessments in that particularformat, I think is incredibly
helpful, because we're movingthings into how we're doing them
(18:03):
in the workplace, which isreally what we should be doing
in any way.
So I think, from that context,that's probably the first thing
that I would be doing.
If we're working more in atrades sort of a college where
we're not as big on computerswe've done a lot of stuff in
hard copy in the past and stufflike that I would have to tread
(18:24):
a lot more slowly with mytrainers.
I would start by doing thingslike just giving them a couple
of brief overview sessions onwhat AI is and what it's capable
of.
I would then probably go into alittle bit of how AI is
impacting their industry.
So like when we think aboutthings like you know brickies
(18:46):
and chippies and stuff like that, how much easier it is to
create plans nowadays, how easyit is to start to use AI for
supporting them withcalculations, adapting plans,
ready, easy access to like hey,what is this building code?
Again, what you know, like maybejust download the GPT app on
(19:08):
their phone and get them tostart to think about things like
that.
I would maybe step in a lotmore and I would focus more on
what is being what, how AI isbeing used in industry, and then
from there, once they start,once that like initial fear is
overcome, I would then start tolockstep them into how they can
(19:28):
use it as a trainer and takethem through that process.
So definitely there would be adifferent process depending on
the type, like the scope of theRTO, and if it was more like
community services, healthservices and things like that, I
would take a similar approach.
But I would also reallystrongly layer that with, like
(19:48):
AI privacy and needing tounderstand how you know how
privacy and protection you knowis different now that we have,
you know, ai happening in thatparticular industry.
Speaker 1 (20:03):
Yeah, yeah.
What do you think could be thechallenges, like you've just
described there, that what you'drecommend, and getting the
trainers on board.
What do you think thechallenges will be when it comes
to resistance?
Speaker 2 (20:17):
Yeah, um, look I, we operate in an industry that is
just so constantly changing like.
We've had new standards thisyear.
We've got new training package,organizing framework with a
completely new structure for ourunits coming out later on this
year, early next year, um, wehave a new avet miss system
(20:42):
being launched and trialed fromnext year.
Our industry is nothing butlike not consistently full of
change, and I think that so manystaff get change fatigue right,
so for them having to look atthis and go, so I'm out of touch
again.
(21:02):
I've got another thing I've gotto learn.
I think that's probably goingto be the biggest hurdle that um
RTOs have to comply with, andthe same as everything.
You've got to take the approachof like look, yes, we've got
more things.
Like, yes, there's more thingswe need to consider.
But how can we start to use AIto make your jobs easier?
(21:24):
So, like, as a trainer, youknow ai, like accelerator's got
this tool where it recommendsfeedback.
Right, fantastic tool fortrainers just to like rather
than having to type out the samething they've only ever typed
out the same thing.
Good job, well done you know,like, okay, it'll give them.
It'll give them a recommendedresponse to the question that
(21:47):
correlates to the question,right?
So, like, let's, let's talkabout the ways that ai can
actually make it easier for youto develop a session plan or,
like you know, some of theselittle tools where, like you can
click on the button and go giveme a summary of this website or
automatically give me sixquestions that I can ask my
(22:07):
students about this website, youknow, just using a little app
and a little button at the topof the screen, right, um, I
think we kind of have to hookthem using those sorts of things
.
Um, because the compliancerequirements continue to change
and update and, as I said, atthe moment we've had 25,000
(22:29):
qualifications cancelled rightby ASQA in the last couple of
months.
They've been cancelled becausethere was a lack of assessment
evidence.
There was no evidence ofcompetency Over the next five
years.
That is not going to be thecompliance problem.
The compliance problem is notgoing to be.
There is not the assessmentevidence there.
There is going to be assessmentevidence there.
(22:50):
The question is is howfraudulent it's going to be
right and you're going to seequalifications being cancelled
because you know they werefraudulently produced out of a
company in Manila.
Right, they were fraudulentlyproduced out of a company in
Manila.
Right, they were fraudulentlyproduced by a very smart AI tool
that literally was designed topump out 32 versions of the same
(23:14):
assessment, all with minormodifications in them, and then
automatically upload them into alearner management system.
Right, this is 100% where we'reheaded At the moment.
We can cross-check some ofthose things, as I said, like
with video uploads of thestudents and cross-comparing.
Five years from now, thestudent will come in, they will
(23:37):
record a five minute video ofthemselves, they will upload it
into an AI tool, the AI toolwill produce a video of them and
you will have video, audio andall the assessment documentation
to prove that a student knowshow to achieve their Cert 3 in
bricklading, how to achievetheir Cert 3.
So for the bad actors out there,right, they are going to have
(24:02):
tools that make it very, verydifficult to detect that they're
doing the wrong thing, and partof the challenge for the
regulator and for those of us inindustry that want to see our
industry do well is going to befiguring out how we catch them.
Yeah, right, and that is goingto be the big, big challenge.
(24:23):
It's not going to be a matterof going into an rto.
You don't have the assessmentevidence and therefore they are
going to have, for the nextthree years, they're going to
have all of the written evidencethat it looks like they've done
the right thing.
Yeah, right, but five yearsfrom now they are going to have
the video and audio evidence tosupport it as well.
It's going to be quick.
(24:44):
It's going to be commerciallyviable, right.
Speaker 1 (24:46):
Yeah, I don't even think it's that far away.
Speaker 2 (24:49):
We really need to think about, like how we protect
our industry and how we can useAI right in the same way.
So, like those automated, likeyou know, a video that turns on
and then takes a live screen.
You know, live screenshots, as,as we're regularly going
(25:09):
through something, you know,validation software, whereby
when the video is taken, um itautomatically, you know, runs a
software program that sitsbehind it to make sure that no
AI, um, you know software isbeing overwhelmed, but it's
going to totally be a cat andmouse game, you know, as it
always has been, but that gameis going to get faster and
(25:31):
faster and faster.
So we have to embrace AI sothat we can stay ahead of the
few bad actors, like the 3% ofbad actors in our industry that
make the 97% of the good guys,you know.
Yeah, drag us down.
That drag them down and thatmake the 97% of the good guys,
you know, yeah, drag us down.
That drag them down and thatmake their lives so much harder,
and so that we can continue toeducate students and go.
(25:51):
If you're not getting the skills, don't go and do it, don't go
and do it, don't pay for thatqualification.
It'll get taken away from you.
You know, and at least now wecan turn around and say 25,000
quals, 25,000 people have paidthree to four thousand dollars
for a qualification, only tohave it taken away from them.
Don't be a sap like that.
Go and do the qualification,get the actual skills, because
(26:15):
that's what's going to lead youinto the industry that actually
allows you to earn the money.
Just you know to do to do thething, yeah, yeah.
Speaker 1 (26:21):
And the thing is you don't want to be turning up to
work where you're expected tomeet those skills levels that
were apparently taught to youand you're not able to do that
because you know you're puttingliabilities on yourself, but
also the RTO, when it comes todoing that as well.
And I think one of the reallyimportant parts and some of the
(26:45):
things that you've alreadytouched on if you haven't been
in AI yet, you need to.
It's a must.
As you said, the students arealready using it.
The trainers need to be usingit so that they understand, and
it doesn't matter what industryyou're in, but once again, I do
agree with you Look at how yourindustry can use AI, if they are
(27:10):
.
If they're not, how could theyuse it?
And then integrate that intohow you're working with the
students.
And I love that idea of gettingthe chat back from.
What prompts are they puttinginto ChatGPT and what are they
getting out and how much do theyactually understand?
(27:32):
And the big thing is theworkforce.
This is the workforce.
Now I'm on ChatGPT all day,every day, like it's my personal
assistant.
Speaker 2 (27:42):
You would never well, I mean you would never ask a
staff member if you were like,hey, I need you to write this
document, this email, this,whatever, you would 100% expect
them.
You actually don't want themsitting there and doing it all
manually.
You're like no, I would ratherpay you half an hour to do the
job using AI than pay you threehours to do the job annually.
(28:04):
Like you and I arebusinesswomen.
Speaker 1 (28:07):
We have a team account, ChatGPT team account,
and all of our team have accessto ChatGPT 100% yeah 100%.
Yeah, yeah, and I keenlyencourage it.
We had a new team member joinus recently who had never even
really used ChatGPT at all, andI actually spent some time
(28:28):
training her on how to use itand she was like, wow, this is
just going to be amazing, it'sjust going to make my work so
much easier.
And I say to her go use ChatGPT, don't spend hours doing it the
old way.
Speaker 2 (28:45):
But use it as a tool.
It's not your, it's not a beall and all solution, it's a
tool.
And so, like I know I'mencouraging my like I
continually go back to my staffand go, hey, like I've noticed,
with five it is far more proneto telling you what you want to
hear as opposed to what you'veasked it for.
So you know, like we go back tothe guys and go, okay, but like
(29:07):
I've got to go back to the botand go, okay, but where did you
get that information?
Okay, now give me the link.
Okay, now, actually, let's goand do this and like, sometimes
I'll like three or four times ina row it was like yes, the new,
it goes against the new rtostandard 5.3.
I'm like there is no newstandard 5.3, dude, like where
are you coming up with this shit?
And they would be like, ohsorry, yes, you are correct.
(29:28):
I'm like, stop driving to me,stop it.
I get pretty upset at it attimes.
Speaker 1 (29:34):
I've had the same thing as like and we've had so
one of my team members.
He uses ChatGPT for our socialmedia and he's not checking what
it's putting out and isreferencing the wrong standards.
I had to go back to him thisweek and say, hey, it's happened
again.
Reminder reminder humaninteraction.
Speaker 2 (29:56):
Human interaction required.
Speaker 1 (29:58):
Yeah, yeah, you can't just have it spit out, and I'd
rather he just take the standardout and just talk about what is
compliance.
He just take the standard outand just talk about what is
compliance, and not put thestandard in there but go back to
the bot.
Are you sure?
Is this the 2025 standards?
Speaker 2 (30:14):
Yeah, and, as we said , like at the end of the day,
it's about the person that'ssitting behind it using the tool
, knowing what they need to know, right.
That's right, like you and me,you know if we'd seen that you
go to me and you go, no, likeyou know, when the bot turns
around and says to me, oh yes,this is against standard 5.3,
I'm like, no, no, it is not.
(30:34):
Because I know there is nostandard 5.3.
And you know you have to havethat person behind it that knows
what they're doing.
So if you're an RTO and youwant to use a compliance tool,
an AI compliance tool, that'sfine, that's your call.
But please go and talk to theperson that sits behind it.
Talk to them for five, 10minutes and see if they actually
have any knowledge of thestandards or any RTO experience
(30:56):
whatsoever.
That would be my onlysuggestion.
Speaker 1 (30:58):
Yes, yeah, yeah, yeah , yes, yeah, yeah, yeah.
And maybe the way you couldtest that is go in with some
compliance questions about maybethings, how you're doing your
practices now and so you canidentify is it actually going to
help you and is it going tostreamline, or are you just
purchasing a software that'sreally not going to do anything?
Speaker 2 (31:24):
No, you know well you know.
We don't currently send ourassessors out to do direct
observations.
We use everything based onthird-party reports.
Is that fully compliant?
And if the RTO bot goes yes,I'm sure there's absolutely no
problems with that then maybeyou need to think twice about
whether or not that's thecorrect tool for you.
Speaker 1 (31:43):
Well, this has been fantastic.
Um, so, just to finish off, ifyou could give rto leaders one
piece of advice when it comes toadopting ai, what would that be
?
Speaker 2 (31:54):
use it as a tool, but only with a human operating it
and making the final call at theend of the day yeah, because we
, uh, we definitely don't wantStarnet to be taking over like
the Terminator.
Yeah, yeah and this is why Ilive in Far North Queensland
(32:15):
because you know we'll be thelast ones to be taken over A bit
like COVID.
I'd like to jump on a boatbefore that happens.
Speaker 1 (32:22):
Yeah Well, thank you very much, Lauren, for once
again joining me on the RTOSuperhero Podcast.
It's always wonderful to have achat with you and talking about
our favourite subject, AI.
Thank you very much and I lookforward to catching up with you
again soon.
Speaker 2 (32:39):
Thanks, Ange, and thank you very much for putting
out this podcast.
Speaker 1 (32:42):
Thank you Awesome, great work.
with me, Angela Connell-Richards, and once again
with Lauren, who's joining mein a great episode, one of our
favourite topics, which is AI.
Welcome Lauren, how are you?
Speaker 2 (00:16):
Good, good.
Thank you so much for having meagain.
Speaker 1 (00:19):
Yes, it's awesome.
So we were chatting offlineabout this and about AI and
where the focus is going in thefuture, in particular, when it
comes to how ASQA are going toaudit us and the different
direction that businesses aregoing in due to AI and that type
(00:43):
of chat, gpt and technology, inparticular, resources and
things like that.
So tell me what drew you to AIfrom the beginning and, in
particular, how it applies inthe LTO sector.
Speaker 2 (00:58):
Yeah, so I did my first talk on AI in 2020, I
think it was 2023, velk was thefirst like public speech that I
did about that.
Obviously, we've been using AIlike GPT, claw, bard, grok,
things like that for that sameperiod of time, watching it,
(01:21):
watching its kind of itsevolution.
And then I was at a conferencea couple of weeks ago, the VETQI
conference and certainly therewas a huge.
You know, there was a lot ofquestions about it, topics about
it, and I think when we haveVELG this year, the VELG
conference I think we're goingto see a lot of speakers talking
(01:42):
about that.
I think there's going to be alot of questions and you know
we've got these new standards inwhich have a really strong
focus on governance, and when Istart to look at how RTOs are
engaging with AI, how trainersand assessors are engaging with
AI, how students are engagingwith AI, I just think it's one
(02:06):
of those spaces that you know issubstantially going to change
how we work.
You know I homeschool my childand you know I had Zach read a
book last week and I said to himokay, I said I want you to sit
down and chat GPT, I want you tokind of you know, create a
structure for a book report.
(02:28):
I want you to cover X, y and Zand then you know you'll need to
work with it to refine thereport.
This is a kid that's got youknow dyslexia and you know hates
writing.
But you know that's a reallyeasy way for him to be able to
interact so that he can create aproper report and effectively
that's how he would create anysort of report or document in
(02:49):
the workplace, right, yeah?
So, you know, we're now, then,starting to look at how
certainly we're having to lookat how we adapt our assessment
tools.
Um, so, in Annawire we build alot of our assessments in
Accelerate for our clients andwe've just redesigned I think
over the last two months we'vehad to rebuild something like
(03:11):
700 or 800 assessments in ourclient systems because we're now
starting to incorporate a livevideo component.
So when we do our validations,we can validate what the student
is typing, like their writtenlanguage, in comparison with
what they can produce verballyand how they can talk to a topic
(03:34):
verbally.
When you can see a student whois writing very academically but
(03:55):
when they speak, their orallanguage would demonstrate an
AQF level two to three levelsbelow what they're talking to
right, same questions, but thewritten version and the oral
versions just aren't lining up.
And that's for us, that'ssomething that we're having to
start to do, because really,when it comes down to validation
, it boils down to okay, but ifyou're going to walk the walk,
(04:19):
you've got to talk the talk aswell, and I think this is
something we're going to startto see Now.
That easily then starts tobring in things like well, we
had to adapt our assessmentinstructions to go.
A live camera is going to comeon and it's going to record you
and please make sure there'snobody in the background and
please make sure you're dressedand professional in your
(04:41):
presentation, and you know.
So we then start to think about, from a privacy perspective,
what can and can't we do.
I know when I was at theconference I was talking to
Accelerate and they had a clientcome to them and go look, can
we create a software where wecan record in the childcare
centres and it automaticallyblows out all of the children's
(05:03):
faces and identities, carecentres and it automatically
blows out all of the children'sfaces and identities?
You know.
So AI is impacting all of thisright and there is so many
different ways in which itaffects people.
You know, as you know, mylittle AI Fireflies follows me
around to all of my meetings andin most of them she's allowed
to participate.
(05:23):
But certainly in any of mygovernment meetings I just turn
around to him, go look, I've gotno choice.
She follows me everywhere.
Just kick her out, you know.
But it's, it's a fantastic toolbecause you know now, when I
used to struggle to todemonstrate all the quality
activities I was doing withclients.
Now, every meeting is recorded,meeting minutes are produced
(05:46):
and then my amazing operationsmanager, laura, takes every
single one of those meetingnotes and copies it into the
client's quality register.
And so whenever we're going toask for and saying what are the
quality activities you're doing,we've got know notes upon notes
upon notes upon notes showingthat every couple of weeks we're
(06:08):
going, we're make, we'retalking about an improvement
we're making or we'reimplementing a plan, and it's
all there and it's so easy, youknow.
So I think what ai could do isabsolutely brilliant in our
sector.
Um, it certainly makes iteasier for me to make far more
innovative and enjoyable andengaging assessments and content
(06:29):
.
But there's a whole other sideof it as well, and you know we
were talking about before.
You know like you've got an AIproduct.
There's three or four of theseAI compliance hubs out there
from, you know, little teenyboppers no offense to the teeny
boppers out there, but you knowlike the photos look like photos
(06:49):
of my 18 year old.
You know who's clearly thislittle group of boys that have
got this great AI knowledge, butnow they're saying that they
can help you run your wholebusiness, but they've never
actually worked in vet beforeyou know.
So there's a lot of stuff outthere that RTOs are now going to
get kind of oh well, you coulddo this and you can do this.
(07:11):
And it's that filtering process.
Again, you know our kidsgrowing up nowadays they don't
have to.
It's not about whether or notthey've got access to the
information, it's filtering theinformation.
What's the quality of it andwhere does it come from?
And I think RTOs have to bejust as savvy.
What is the tool that I'mlooking at?
(07:32):
What is it built on?
Who is it coming from?
Right, and, and is thisactually something that's going
to save my business money andgive me a return on investment?
Or, you know, am I just buyingsomething that I think is going
to tick a box when really it'snot actually going to make any
(07:54):
sort of measurable impact on mybusiness?
Speaker 1 (07:55):
Yes, so what do you think are the biggest risks for
RTOs when it comes to adoptingAI?
Speaker 2 (08:03):
So I was having a really interesting conversation
with Vanessa, from Prickly toSweet, who's like she's been in
AI for what?
Eight or nine years now, andthen I think Kerry Buttery as
well, is another one that I'vehad a couple of chats with, and
she's very much operating inthat AI space too.
Where your data goes issomething that I've taken from
(08:24):
both of them.
Like that's definitelysomething what data you're being
asked to put into the AI andwhat data you're getting out of
it, and how it then secures it.
Is it being held in an offshoreserver or an Australian server?
Are there any?
That's the biggest question.
Speaker 1 (08:38):
I get asked.
Speaker 2 (08:40):
Yeah.
Speaker 1 (08:41):
When I refer people to Accelerate.
Speaker 2 (08:44):
I know they've got all all like.
They've got 27 different.
You know security controls intheir system, um, but you know
like and I mean I, you know thegovernment's using microsoft, so
microsoft and microsoftco-pilot and stuff like that
obviously their security andthings like that is approved,
but there's a lot of others that, like, trying to get that
information from them isproblematic, you know.
(09:07):
So I think that's one thingRTOs need to start to learn to
navigate and negotiate.
And then the Geico principle ofyou know people who have built
AI systems that don't have RTOknowledge, that can't
(09:29):
demonstrate that they've got RTOknowledge.
And you know not justnecessarily even RTO knowledge.
If you're looking at somethingto run your whole business and
your business runs six fundingcontracts, you know, okay, I can
write a policy set for your RTOstandards, but there is not a
single funding contract thatdoesn't require you to adapt
(09:50):
your policies for that fundingcontract.
You know, in terms of finance,in terms of how you capture your
data, in terms of yourenrolment process, in terms of
additional, you know things thatyou need to capture from the
students or information you needto provide to the students and
things like that.
So I just think that peopleneed to learn how to have really
(10:12):
good conversations with anyprovider that they're going to
look to use any sort of resourcefrom, and I would be super, I
would be very reluctant.
Um, and I've seen this, I'veseen this advertised on a few of
these RTO, um assessmentdevelopment things where they go
use this tool and you'll, wecan build, you'll, it'll, it'll
(10:35):
build your resources in fiveminutes.
It'll build a compliantcontextualized assessment in
five minutes.
And I'm just going to turnaround to people and say it's
bullshit.
Yeah, I have not seen a systemthat can write an assessment
tool in five minutes that'scompliant.
Most of them are built off ofgpt, which is built off of the
internet, and the internet isfull of non-compliant assessment
(10:57):
tools.
So you know, if it sounds toogood to be true, it's probably
too good to be true.
Speaker 1 (11:05):
So with ComplyHub so you've touched on that the
system that we've built, we'veactually integrated it with TGA
so it will actually bring in theinformation for the training
products that are on your scopeof registration.
So it does make a bigdifference.
But you know, even just youknow, having the unit of
(11:25):
competency and downloading thatfrom TGA and putting that into
chat GPT, you're not going toget a good assessment tool
because it's just going tofollow the legislation or the
requirements of the trainingproduct.
It's not going to go intoreal-world scenarios, it's going
to create an assessment againstPC 1.1 and one against PC 1.2.
Speaker 2 (11:46):
And you know what you might actually get like within
the next year.
You'll probably get maybe acompliant-ish tool, but it's not
going to be usable.
The assessors are going to hateit.
It's going to be overly, it'sgoing to probably be far more
academic than it should be forthe AQF level.
Like there's so many nuances,we still need humans Still need
(12:07):
humans.
Yes, thank you.
Yes, it's a great tool.
It's an great tool.
It's an amazing tool and I useit every day and my business
would be a lot more cumbersomeand more expensive without it.
I love it.
I think it's amazing what it'sgoing to do, but there has to be
that human component behind it,and anyone purchasing a service
(12:30):
that uses AI needs to talk tothe people that sit behind it.
So, like, if you're going to goand use Vanessa's AI validator,
go talk to Vanessa and see theknowledge that she has.
If they're going to use your AIcompliance hub, go and talk to
you and your team and see thecompliance that sits behind it.
I know that.
You know the learning resourcesgroup has got a couple of
(12:51):
different AI tools that they'rebringing out.
You know they've been in thisindustry and they've actually
built good, good assessmenttools and this is coming from
somebody who builds assessmenttools.
But, like we've always had a lotof respect for the learning
resources group because you knowthey don't sell these very
cheap assessments, you know thatthe contract requires you to.
(13:12):
You know, take seven stepsbefore you get any sort of
compliance guarantee out of themand things like that you know
they genuinely mean well, theygenuinely want to provide a
really good quality product, andyou know you need to talk to
the people that sit behind itand ask those questions, and the
(13:34):
governance standards are veryclear.
Do your due diligence.
Know the people that you'reworking with.
Know the companies, know thepeople that sit behind the
companies that you want toassociate yourself with.
Speaker 1 (13:44):
Yeah, yeah definitely , and I think one of the big
things around using AI withinyour like what I see is I need
it as part of what we do now asa service.
However, you still need to havea good understanding of the
requirements of the legislation,the standards, the different
(14:07):
other different requirementsthat you need to have in place
for a range of differentlegislation, and I think someone
trying to develop a compliancesystem who doesn't actually have
any idea about compliance isnot going to be able to build a
good system.
Like I'm now able to build asystem without a software
(14:30):
developer, and it's fantasticbecause I have built stuff with
software developers before.
It was always trying to teachthem what you do and how to do
it, whereas now I don't have todo that.
I'm doing it straight,one-on-one with the AI and I'm
teaching it and it remembers,which is great, and we're
building on that over and overand over again.
(14:52):
So it's, yeah, so much betterthan trying to work with a
software developer who's reallytech savvy but just really
doesn't understand the topicthat you're talking about, and I
totally agree with you therethat you really need to have a
good understanding of who you'reworking with and whoever's
building these systems?
Do they understand what are therequirements when it comes?
(15:16):
Because the last thing you wantis to turn up in an audit and
your auditor picks up somemistake that your AI, that AI
has made, put it at risk,particularly if you're getting
assessment tools built andputting your RTO at risk of
(15:36):
certificates being cancelled dueto not collecting sufficient
evidence.
Yeah, yeah, definitely so I'mgoing to move on to another
question.
For an RTO that hasn't startedyet with any AI, what would you
recommend would be the firstplace they should start?
Speaker 2 (15:58):
I would say it would depend on the scope of the RTO.
So I think for those that areoperating in more of the like
soft services, like business,financial IT, anything of that
sort of space, the first thingthat you need to know is that
your students are 100% using AInow, and if you're not, I can
almost guarantee that yourassessments are being
(16:21):
fraudulently submitted andyou're completely unaware of it.
So my first step in an RTO likethat would be to get the
assessors working with AI veryregularly, even if it's just
helping them to create theirsession plans.
Speaker 1 (16:36):
Contextualizing.
Speaker 2 (16:38):
Yep, contextualizing certain things, maybe you know,
getting three or fourassessments together and working
with AI to do the validation orsomething like that, just so
that they can start tounderstand the way that AI
writes the formatting that itprefers, the spelling that it
prefers, the spelling that itprefers, because that's going to
make it much easier for them tostart to pick up on that with
(16:59):
their students and also get themto start to think about how
they can update theirassessments to incorporate more
of AI.
Like I know, when we're writingpolicies now we're writing
assessments, now we'll dosomething like okay, you need to
write a policy, you need todevelop and implement a policy
on X, right?
What you're going to do isyou're going to work with AI to
(17:20):
develop a template all right anddevelop the draft, and then
you're going to go back andforth until you refine it and
get it to exactly what it is.
You're then going to submit acopy of the chat with your AI
and the finalized document,right?
That is exactly aligned withwhat we would expect in the
workplace.
(17:40):
We're not putting limits onstudents about working with AI,
but we are getting the AItranscript with the students so
that we can see what thestudent's actual knowledge is
right.
Yeah, and teaching trainers howto like, tweak and adapt
assessments in that particularformat, I think is incredibly
helpful, because we're movingthings into how we're doing them
(18:03):
in the workplace, which isreally what we should be doing
in any way.
So I think, from that context,that's probably the first thing
that I would be doing.
If we're working more in atrades sort of a college where
we're not as big on computerswe've done a lot of stuff in
hard copy in the past and stufflike that I would have to tread
(18:24):
a lot more slowly with mytrainers.
I would start by doing thingslike just giving them a couple
of brief overview sessions onwhat AI is and what it's capable
of.
I would then probably go into alittle bit of how AI is
impacting their industry.
So like when we think aboutthings like you know brickies
(18:46):
and chippies and stuff like that, how much easier it is to
create plans nowadays, how easyit is to start to use AI for
supporting them withcalculations, adapting plans,
ready, easy access to like hey,what is this building code?
Again, what you know, like maybejust download the GPT app on
(19:08):
their phone and get them tostart to think about things like
that.
I would maybe step in a lotmore and I would focus more on
what is being what, how AI isbeing used in industry, and then
from there, once they start,once that like initial fear is
overcome, I would then start tolockstep them into how they can
(19:28):
use it as a trainer and takethem through that process.
So definitely there would be adifferent process depending on
the type, like the scope of theRTO, and if it was more like
community services, healthservices and things like that, I
would take a similar approach.
But I would also reallystrongly layer that with, like
(19:48):
AI privacy and needing tounderstand how you know how
privacy and protection you knowis different now that we have,
you know, ai happening in thatparticular industry.
Speaker 1 (20:03):
Yeah, yeah.
What do you think could be thechallenges, like you've just
described there, that what you'drecommend, and getting the
trainers on board.
What do you think thechallenges will be when it comes
to resistance?
Speaker 2 (20:17):
Yeah, um, look I, we operate in an industry that is
just so constantly changing like.
We've had new standards thisyear.
We've got new training package,organizing framework with a
completely new structure for ourunits coming out later on this
year, early next year, um, wehave a new avet miss system
(20:42):
being launched and trialed fromnext year.
Our industry is nothing butlike not consistently full of
change, and I think that so manystaff get change fatigue right,
so for them having to look atthis and go, so I'm out of touch
again.
(21:02):
I've got another thing I've gotto learn.
I think that's probably goingto be the biggest hurdle that um
RTOs have to comply with, andthe same as everything.
You've got to take the approachof like look, yes, we've got
more things.
Like, yes, there's more thingswe need to consider.
But how can we start to use AIto make your jobs easier?
(21:24):
So, like, as a trainer, youknow ai, like accelerator's got
this tool where it recommendsfeedback.
Right, fantastic tool fortrainers just to like rather
than having to type out the samething they've only ever typed
out the same thing.
Good job, well done you know,like, okay, it'll give them.
It'll give them a recommendedresponse to the question that
(21:47):
correlates to the question,right?
So, like, let's, let's talkabout the ways that ai can
actually make it easier for youto develop a session plan or,
like you know, some of theselittle tools where, like you can
click on the button and go giveme a summary of this website or
automatically give me sixquestions that I can ask my
(22:07):
students about this website, youknow, just using a little app
and a little button at the topof the screen, right, um, I
think we kind of have to hookthem using those sorts of things
.
Um, because the compliancerequirements continue to change
and update and, as I said, atthe moment we've had 25,000
(22:29):
qualifications cancelled rightby ASQA in the last couple of
months.
They've been cancelled becausethere was a lack of assessment
evidence.
There was no evidence ofcompetency Over the next five
years.
That is not going to be thecompliance problem.
The compliance problem is notgoing to be.
There is not the assessmentevidence there.
There is going to be assessmentevidence there.
(22:50):
The question is is howfraudulent it's going to be
right and you're going to seequalifications being cancelled
because you know they werefraudulently produced out of a
company in Manila.
Right, they were fraudulentlyproduced out of a company in
Manila.
Right, they were fraudulentlyproduced by a very smart AI tool
that literally was designed topump out 32 versions of the same
(23:14):
assessment, all with minormodifications in them, and then
automatically upload them into alearner management system.
Right, this is 100% where we'reheaded At the moment.
We can cross-check some ofthose things, as I said, like
with video uploads of thestudents and cross-comparing.
Five years from now, thestudent will come in, they will
(23:37):
record a five minute video ofthemselves, they will upload it
into an AI tool, the AI toolwill produce a video of them and
you will have video, audio andall the assessment documentation
to prove that a student knowshow to achieve their Cert 3 in
bricklading, how to achievetheir Cert 3.
So for the bad actors out there,right, they are going to have
(24:02):
tools that make it very, verydifficult to detect that they're
doing the wrong thing, and partof the challenge for the
regulator and for those of us inindustry that want to see our
industry do well is going to befiguring out how we catch them.
Yeah, right, and that is goingto be the big, big challenge.
(24:23):
It's not going to be a matterof going into an rto.
You don't have the assessmentevidence and therefore they are
going to have, for the nextthree years, they're going to
have all of the written evidencethat it looks like they've done
the right thing.
Yeah, right, but five yearsfrom now they are going to have
the video and audio evidence tosupport it as well.
It's going to be quick.
(24:44):
It's going to be commerciallyviable, right.
Speaker 1 (24:46):
Yeah, I don't even think it's that far away.
Speaker 2 (24:49):
We really need to think about, like how we protect
our industry and how we can useAI right in the same way.
So, like those automated, likeyou know, a video that turns on
and then takes a live screen.
You know, live screenshots, as,as we're regularly going
(25:09):
through something, you know,validation software, whereby
when the video is taken, um itautomatically, you know, runs a
software program that sitsbehind it to make sure that no
AI, um, you know software isbeing overwhelmed, but it's
going to totally be a cat andmouse game, you know, as it
always has been, but that gameis going to get faster and
(25:31):
faster and faster.
So we have to embrace AI sothat we can stay ahead of the
few bad actors, like the 3% ofbad actors in our industry that
make the 97% of the good guys,you know.
Yeah, drag us down.
That drag them down and thatmake the 97% of the good guys,
you know, yeah, drag us down.
That drag them down and thatmake their lives so much harder,
and so that we can continue toeducate students and go.
(25:51):
If you're not getting the skills, don't go and do it, don't go
and do it, don't pay for thatqualification.
It'll get taken away from you.
You know, and at least now wecan turn around and say 25,000
quals, 25,000 people have paidthree to four thousand dollars
for a qualification, only tohave it taken away from them.
Don't be a sap like that.
Go and do the qualification,get the actual skills, because
(26:15):
that's what's going to lead youinto the industry that actually
allows you to earn the money.
Just you know to do to do thething, yeah, yeah.
Speaker 1 (26:21):
And the thing is you don't want to be turning up to
work where you're expected tomeet those skills levels that
were apparently taught to youand you're not able to do that
because you know you're puttingliabilities on yourself, but
also the RTO, when it comes todoing that as well.
And I think one of the reallyimportant parts and some of the
(26:45):
things that you've alreadytouched on if you haven't been
in AI yet, you need to.
It's a must.
As you said, the students arealready using it.
The trainers need to be usingit so that they understand, and
it doesn't matter what industryyou're in, but once again, I do
agree with you Look at how yourindustry can use AI, if they are
(27:10):
.
If they're not, how could theyuse it?
And then integrate that intohow you're working with the
students.
And I love that idea of gettingthe chat back from.
What prompts are they puttinginto ChatGPT and what are they
getting out and how much do theyactually understand?
(27:32):
And the big thing is theworkforce.
This is the workforce.
Now I'm on ChatGPT all day,every day, like it's my personal
assistant.
Speaker 2 (27:42):
You would never well, I mean you would never ask a
staff member if you were like,hey, I need you to write this
document, this email, this,whatever, you would 100% expect
them.
You actually don't want themsitting there and doing it all
manually.
You're like no, I would ratherpay you half an hour to do the
job using AI than pay you threehours to do the job annually.
(28:04):
Like you and I arebusinesswomen.
Speaker 1 (28:07):
We have a team account, ChatGPT team account,
and all of our team have accessto ChatGPT 100% yeah 100%.
Yeah, yeah, and I keenlyencourage it.
We had a new team member joinus recently who had never even
really used ChatGPT at all, andI actually spent some time
(28:28):
training her on how to use itand she was like, wow, this is
just going to be amazing, it'sjust going to make my work so
much easier.
And I say to her go use ChatGPT, don't spend hours doing it the
old way.
Speaker 2 (28:45):
But use it as a tool.
It's not your, it's not a beall and all solution, it's a
tool.
And so, like I know I'mencouraging my like I
continually go back to my staffand go, hey, like I've noticed,
with five it is far more proneto telling you what you want to
hear as opposed to what you'veasked it for.
So you know, like we go back tothe guys and go, okay, but like
(29:07):
I've got to go back to the botand go, okay, but where did you
get that information?
Okay, now give me the link.
Okay, now, actually, let's goand do this and like, sometimes
I'll like three or four times ina row it was like yes, the new,
it goes against the new rtostandard 5.3.
I'm like there is no newstandard 5.3, dude, like where
are you coming up with this shit?
And they would be like, ohsorry, yes, you are correct.
(29:28):
I'm like, stop driving to me,stop it.
I get pretty upset at it attimes.
Speaker 1 (29:34):
I've had the same thing as like and we've had so
one of my team members.
He uses ChatGPT for our socialmedia and he's not checking what
it's putting out and isreferencing the wrong standards.
I had to go back to him thisweek and say, hey, it's happened
again.
Reminder reminder humaninteraction.
Speaker 2 (29:56):
Human interaction required.
Speaker 1 (29:58):
Yeah, yeah, you can't just have it spit out, and I'd
rather he just take the standardout and just talk about what is
compliance.
He just take the standard outand just talk about what is
compliance, and not put thestandard in there but go back to
the bot.
Are you sure?
Is this the 2025 standards?
Speaker 2 (30:14):
Yeah, and, as we said , like at the end of the day,
it's about the person that'ssitting behind it using the tool
, knowing what they need to know, right.
That's right, like you and me,you know if we'd seen that you
go to me and you go, no, likeyou know, when the bot turns
around and says to me, oh yes,this is against standard 5.3,
I'm like, no, no, it is not.
(30:34):
Because I know there is nostandard 5.3.
And you know you have to havethat person behind it that knows
what they're doing.
So if you're an RTO and youwant to use a compliance tool,
an AI compliance tool, that'sfine, that's your call.
But please go and talk to theperson that sits behind it.
Talk to them for five, 10minutes and see if they actually
have any knowledge of thestandards or any RTO experience
(30:56):
whatsoever.
That would be my onlysuggestion.
Speaker 1 (30:58):
Yes, yeah, yeah, yeah , yes, yeah, yeah, yeah.
And maybe the way you couldtest that is go in with some
compliance questions about maybethings, how you're doing your
practices now and so you canidentify is it actually going to
help you and is it going tostreamline, or are you just
purchasing a software that'sreally not going to do anything?
Speaker 2 (31:24):
No, you know well you know.
We don't currently send ourassessors out to do direct
observations.
We use everything based onthird-party reports.
Is that fully compliant?
And if the RTO bot goes yes,I'm sure there's absolutely no
problems with that then maybeyou need to think twice about
whether or not that's thecorrect tool for you.
Speaker 1 (31:43):
Well, this has been fantastic.
Um, so, just to finish off, ifyou could give rto leaders one
piece of advice when it comes toadopting ai, what would that be
?
Speaker 2 (31:54):
use it as a tool, but only with a human operating it
and making the final call at theend of the day yeah, because we
, uh, we definitely don't wantStarnet to be taking over like
the Terminator.
Yeah, yeah and this is why Ilive in Far North Queensland
(32:15):
because you know we'll be thelast ones to be taken over A bit
like COVID.
I'd like to jump on a boatbefore that happens.
Speaker 1 (32:22):
Yeah Well, thank you very much, Lauren, for once
again joining me on the RTOSuperhero Podcast.
It's always wonderful to have achat with you and talking about
our favourite subject, AI.
Thank you very much and I lookforward to catching up with you
again soon.
Speaker 2 (32:39):
Thanks, Ange, and thank you very much for putting
out this podcast.
Speaker 1 (32:42):
Thank you Awesome, great work.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!