In this episode of PLATO Panel Talks, host Mike Hrycyk is joined by Jesslyn Dymond (TELUS) and Jason Kicknosway (PLATO) for a deep dive into what it really takes to build trustworthy, responsible AI systems. Together, the panel discusses their hands-on collaboration using Purple Teaming—a powerful approach that combines offensive (Red Team) and defensive (Blue Team) strategies to test, break, and strengthen AI before it reaches the public.
This conversation goes beyond the hype of AI and into the reality of implementation. Jason shares what it was like for PLATO testers to step into AI testing for the first time—learning prompt engineering, experimenting across tools, and even testing voice systems—and how these efforts revealed how sensitive AI can be to nuance, and how critical diverse perspectives are when evaluating performance.
Jesslyn brings a governance lens to the discussion. With only a fraction of Canadians confident in how companies use AI, she highlights the growing need for transparency, explainability, and inclusive design—especially the integration of Indigenous perspectives that are often missing from mainstream AI models.
Ultimately, this episode is about more than technology. It’s about collaboration, co-creation, and ensuring AI enhances human experiences rather than replacing them. Whether you’re just starting your AI journey or already building with it, this conversation offers a grounded look at how to approach AI responsibly—by testing deeply, thinking broadly, and bringing more voices into the process.
Having trouble with the player?
Listen to this episode on Spotify (opens in a new tab)
This option provides a more reliable playback experience on mobile devices.
Episode Transcript:
Mike Hrycyk (00:00):
Hello everyone. Welcome to another episode of PLATO Panel Talks. I’m your host, Mike Hrycyk, and today we’re going to talk about AI. This year we’ve decided that we’re going to do a lot of focus on AI, and we have different topics with different discussion points. But with this one, I thought it would be really interesting to talk about a project where AI is included, and it’s hands-on, and we’re actually doing the implementation. And to match up with that, I brought in a couple of experts who were part of that project. So, I’m going to turn it over to you, Jesslyn, to introduce yourself.
Jesslyn Dymond (00:26):
Hi everybody. My name is Jesslyn Dymond, and I have the privilege of leading AI governance at TELUS. So, TELUS is a partner of PLATO, and we’ve been working together to enhance our testing and explore what’s possible with AI as we learn how to manage it. It’s an exciting time in the world of AI. And I tell my kids my job is keeping the robots at bay, and in partnership with PLATO, we’ve developed some strategies to bring that to life.
Mike Hrycyk (00:58):
Thank you for joining us. It reminds me of herding cats. Herding the robots is like herding cats.
Jesslyn Dymond (01:03):
Less cute, but for sure.
Mike Hrycyk (01:05):
Jason, tell us about yourself.
Jason Kicknosway (01:07):
Hello everyone. My name is Jason Kicknosway. I work here at PLATO. I am a quality assurance software tester. I’m an apprentice after coming through the program that PLATO offered to train software testers.
Mike Hrycyk (01:19):
Great. Thanks, Jason. Thanks for joining us. So, let’s start. I love starting with level-setting in my podcasts. Let’s start with you, Jesslyn, and tell us about the project that you guys brought PLATO in as a partner for.
Jesslyn Dymond (01:31):
I’d love to level-set a bit more generally and share a little bit about the work that we do in AI. So, TELUS is definitely a significant telecommunications phone company in Canada, but we also do lots of work in healthcare, agricultural technology, and much of that is powered by AI. And when we’re using AI, we want to find ways to make sure it works as accurately and as responsibly as possible. And when we talk about responsible AI, that means ensuring that it works in a way that is consistent with people’s expectations. It isn’t going to be surprising or say something unexpected.
(02:14):
And if you’re learning about AI or using it, that is definitely something that is going to happen. So, when we put our AI systems out there into the world, we go through a lot of reviews, and that’s part of what my team does: help implement and develop solutions, or something we call guardrails, to make sure they work as expected.
(02:37):
When we started using large language models, which is the type of generative AI that’s very conversational and is widely available now in so many different platforms, kind of like powering chatbots. The first time we built a system that was going to be available for the public, we went through a very thorough review of what was available and how to use it. And one of the concepts that came out of the technical documentation from the company that developed this AI model was that there should be Red Teaming of the system. So, I said, “What is that? ” So, I did a little Google, “What is Red Teaming?” We didn’t have the same AI tools available at that time. And Red Teaming is a technique in cybersecurity where you try to attack and break the system from the inside before you release it.
(03:27):
So, I talked to my friends in cybersecurity and learned about Red Teaming. And they’re like, “You know what’s better though, Jocelyn, is when you also do Blue Teaming, which is trying to build guardrails and ensure and fix those gaps that are being discovered.” But the ultimate solution is to do Purple Teaming, and that is bringing the Red Team and the Blue Team together and working collaboratively to find those problems, fix them, and ensure that the system is as robust as possible.
(03:56):
Well, this got me really excited because when I work with AI, I want to make sure that we are including as many people as possible in ensuring that it works well. And so, we set up a Purple Team at TELUS where we invited volunteers from all over the company to come and help us break the system and work with the developers to fix the system. And we even get ideas about how to fix the system from people in finance, HR. And our developers were like, “Whoa, whoa, whoa. We did our quality assurance. We’re good, Jocelyn.” But what we hadn’t really anticipated was the diversity of ways that a system could be used when you make it really widely available. And what’s super interesting with these generative AI systems is the unpredictability of their outcomes, as there are far too many bad examples of.
(04:40):
And so, we wanted to push it to the limits. We launched this Purple Team, and it became the norm in how we developed AI solutions because it was extremely helpful. There was so much that we learned and really discovered the value of having a broad range of perspectives and expertise, different mindsets, all looking at a tool together.
(05:04):
So, when I met Denis Carignan, one of the leaders at PLATO as part of the work that TELUS does in reconciliation, we have a reconciliation action plan, and heard about the work PLATO was doing in both software testing – something I’m learning very much about because it’s not typical that we do this type of testing in our work. But also, in building out the Indigenous IT workforce and the apprenticeship and training programs that are available through PLATO, I thought, wow, it would be so interesting to learn more about how we could benefit from that testing expertise. But also extend the work we’re doing in Purple Teaming, which we’ve done in a variety of settings over the past couple of years, including with the public, but to have the opportunity to include Indigenous expertise in that exercise of testing and exploring what happens when you prompt systems or present different queries and have different results.
(05:59):
And so, TELUS has this commitment to incorporate Indigenous perspectives into our data ethics and AI strategies. And when you look at the type of work PLATO is doing in software testing and the opportunities that are out there with AI to really transform the way technology works, the way testing happens. And most importantly, I would say from my seat as the person trying to figure out how do we keep these robots from taking over, it’s really about ensuring we have a wide range of voices at the table in the development of this technology. And it was really exciting to learn about how we could partner together to make this happen. So, that’s the story of how we set up Purple Teaming and where the partnership with PLATO came in.
Mike Hrycyk (06:46):
Awesome. I had not heard of Red, Blue and Purple Teaming. And I really like that it fits with the colour wheel. Purple comes out of red and blue mixed. I really like that. That’s descriptive. Guardrail testing, we don’t term it that often that way in QA, but it has always been a big thing in QA. It’s not just as simple as finding, “Hey, is there a bug?” It’s when a bug happens – because eventually they will always happen – have we written a system that is robust and user-friendly enough to give you a result that helps the user – it’s not a crash. It’s not a blue screen. It’s what does happen when a bug happens? And that’s part of writing a bug.
(07:21):
So, you might write a bug that said, “Oh yeah, when you do this and this and enter this, then the system doesn’t work.” But then you might write a second bug that says, “But it’s really user-unfriendly. You don’t say what the user should do. You don’t tell them where they should go for help.” And so, that’s guardrail. And then I think this is an extension of that idea.
Jesslyn Dymond (07:37):
There’s so much need for that kind of explainability, another concept in AI, or transparency, of helping people understand what it’s doing. And the reasoning technology that we’re seeing with AI is kind of helping where you see it process and explain, “I’m doing this, I’m doing that. ” But the trust is what we need to build out, and it’s not until you show the capabilities and where there are strengths and where there are limitations that you can really build that trust.
Mike Hrycyk (08:08):
Alright. Jason, why don’t you tell us a bit about the project that you were brought into and what it does and how you and the team interacted. What did the team look like?
Jason Kicknosway (08:16):
Working on the TELUS project, it was my first full foray into using AI. Before I would use it just like Google, like I would ask it a simple question, it would give me the response I needed. But jumping into the TELUS project, it forced our whole team to start learning AI. So, we went to pretty much every single AI you can think of, from Claud, ChatGPT, and Copilot. We were pulling all of that information in and trying to learn how to use AI. And then once we learned how to use AI, we had to incorporate it into a testing framework.
(08:50):
So, that’s where we started learning about prompt engineering. So, we wanted to ensure that when we produced a result, we wanted the exact result that we were kind of looking for. And that’s all included into prompt engineering where everything is literal when you talk to a computer. So, changing one word inside of your prompt changes the outcome that comes out. And that was something that was a little bit different for us to learn because one simple little change had a different output. But once we figured everything out and we were 100% Purple Teaming, we were doing both validation prompts, and we were doing some of the other more destructive prompts. So, we were looking for biases, hallucinations, and misinterpretations of what we were typing into our prompts.
(09:39):
So, in the first half, we worked on prompt engineering. So, we were going through the website. We asked the AI attached to the website how to find certain things on each website. So, each separate little province, all the way down to both French, and we’ve actually incorporated some Indigenous language into some of our prompts to see what would come out of the responses. So, it was a very, very educational experience working on this project.
Jesslyn Dymond (10:10):
So, we use AI in a number of different ways, but it’s really important that it’s handled on a use-case-by-use-case basis because those guardrails need to be tuned for each specific implementation. So, one of the systems Jason is describing is our customer service AI tool. So, that’s available for anyone to use on telus.com, and it helps you find products, services, and support documentation. There are other implementations that are available only to customers that are authenticated. A little bit lower risk because it’s not widely available on the internet. But in that case, we can be dealing with implementations that are far more sensitive, like helping someone find a relevant piece of support documentation to analyze the symptoms of a patient that they’re working with or finding the right network topology document to support a technician in the field trying to fix a router that needs some configuration changes.
(11:08):
So, in each of those contexts, the testing in the very beginning needs to be focusing on making sure that it’s staying on topic, right? Why would you want your network bot to be telling you about lines from a TV show or talking like a pirate, as is everyone’s favourite joke to play on the AI? And in the context of a customer interaction that could be harmful or offensive, and jokes are always where things go off the rails with gen AI systems.
Mike Hrycyk (11:35):
Excellent. Anything helping us get through the customer service portal to where we need to be is awesome. And as your services grow and grow and grow, it becomes harder and harder, right? So, AI is definitely something that can help with that.
Jesslyn Dymond (11:48):
That’s right. And we want to really manifest that opportunity. That’s why the role of the team that I’m part of, TELUS’s Data and Trust Office, is so critical because we do research with Canadians to understand where they’re at. And only 34% of Canadians in our recent 2026 AI report indicate that they trust companies using AI. And when you read some of the headlines, it’s not surprising that that is the case. We need to help people understand how the technology works and why they can trust it. To do that, it needs to work consistently and to help provide the expected results. There’s nothing more frustrating than being stuck talking to AI when you need the support to move forward, and having that kind of human-centric support and the right testing in place is what provides the assurance that the systems are going to work in a way that’s ultimately delivering on the promise of having all of this technology work better and for our own good.
Mike Hrycyk (12:54):
Yeah. If the trust is only, did you say 34% of people trust? If the trust is only at 34%, and yet 100% of executive is telling people to use AI –
Jesslyn Dymond (13:06):
We have our work cut out for us.
Mike Hrycyk (13:09):
Yeah. I mean, every person who has an interaction that they can tie to AI needs to have a positive reaction or else the trust won’t go up, right? But it’s going to be there because there are so many benefits.
(13:21):
So, Jason, what do you think – the team that we put together for this project, what do you think they brought to this engagement that was unique and special?
Jason Kicknosway (13:30):
Wow. We had people who were more community-based on our team, and on the far spectrum we had very program-based people on our team. It was great to combine both of those worlds together. So, we had the technical side, talking with our community side, and that helped us formulate how we were going to test the AI. So, it’s always a good measure to have a well-balanced team, and I think our team was pretty well-balanced.
Mike Hrycyk (13:58):
So, Jesslyn, I think you’ve answered my question around why you brought us in and what were our goals, but let’s extend that a little bit. How well do you think the team succeeded in helping you attain your goals? And the tough question, were you able to measure our success?
Jesslyn Dymond (14:15):
Indeed, yes. The team was tremendously helpful and actually kind of enhanced – I shouldn’t be surprised – but the procedures we had in place. So, you have a team that’s learning how to use AI and some of the techniques around adversarial testing and trying to trick and make the system perform at its limit. And then my office, which is full of folks who are AI experts, but don’t have the same expertise or training or the experience of being apprentices in a software testing training program. So, we learned so much about how to standardize our processes of testing and create those measurements.
(14:56):
So, something like hundreds of copilots. A copilot is a function that we have that team members can use to create their own little AI system and share it with their team to analyze internal documents or create weather reports for the areas that they’re supporting. And there was all of these copilots that have been built ad hoc, and the PLATO team interrogated them and created standardized outputs that we’re able to continue to incorporate into our procedures for Purple Teaming. And so, that’s now part of the way that we do our testing. And having that rigour and that methodology that we were not familiar with was critically important.
(15:39):
In addition to the community perspective that Jason mentioned, a big part of why we want to work with an Indigenous-owned company like PLATO is to hear from the voices of Indigenous peoples across Canada. And so, there were use cases where call transcripts, like those recordings of interactions with a customer service account or healthcare professional were enhanced and really supported to meet the needs of people who are Indigenous across Canada and have that perspective to bring a realistic or appropriate voice to help the model, the AI system, provide valid inputs and shape the way that these sorts of conversations or interactions happen.
(16:20):
And that is a really, really important piece of knowledge or a way of understanding and learning. This breadth of humanity that is out there is not what is captured by default in the inherently trained on the internet large language models that are very widely used. It’s a really, really narrow set of perspectives from a pretty tight geography or slice of the internet. And so, we want to find ways to have that enhanced and to recognize that right now we need tremendous human oversight, and that’s the expectation that we’re hearing from our customers. They don’t want these systems making decisions or providing advice without an expert overseeing it. But to really bring about some of these benefits and unlock the real potential, we need to have more trustworthy mechanisms that can help provide an expected response, an accurate response and unlock some of the additional productivity benefits to build in this type of control by default where we don’t have to have humans validating every single outcome.
Mike Hrycyk (17:25):
Something you said in there made a lightbulb of brilliance go off for me. One of the things that is panned nowadays, especially in technology is small talk. People hate small talk, but the benefits of small talk are immense and huge, especially if you’re going to have a very short-term interaction, like a customer service interaction, where you have to get something quickly, but you want it to stay as positive as possible – and people are coming into these things irate. Being able to connect on a little tiny level at the start of that makes the interaction more human, as you said. And the brilliance part, the light bulb was the idea that your reps had developed this idea of having little copilot interactions that would look up the weather where the person is calling from, right? Because the weather is the number one starting point for small talk, people are passionate about their weather. We started this conversation talking about the snow you’re having in Toronto, and it’s just, that’s so smart. Don’t start today with AI doing everything. Start today with AI helping, helping you be a better human. I like that. AI should help you be a better human.
Jason Kicknosway (18:25):
I always say you should treat AI like your personal assistant, a friend that you take everywhere with you. I mean, it’s in your phone, so definitely use AI.
Mike Hrycyk (18:34):
I’ve been seeing and reading different things where some people are saying, “Make sure you treat your AI like a robot in your interactions.” And then there’s a bunch of people saying, “Make sure you treat AI politely because you get different interactions.” I just can’t stop from being polite. I say please and thank you and stuff, and it works. And so, I don’t know. I don’t know what the right answer is.
Jesslyn Dymond (18:53):
And the behavioural impacts of that, right? I’m not a fan of telling people not to use their manners. There’s certainly a tension, and I have gone back and forth on the concept of anthromorphization, or do we want to humanize AI, or should it always be something that is an assistant? And we try to really recognize that it is uniquely artificial. When we talk about intelligence or artificial intelligence, as the field is so objectively named, the intelligent part is us. That is where the real expertise is going to remain. It is an aspiration to have artificial intelligence, but is it really?
(19:33):
I think we know so well how to relate to each other, and that is part of what makes us human, and what we need to continue to respect, uphold, and be conscious and intentional about where we bring AI in. And that’s going to be something we figure out together. And as I say all the time when we’re Purple Teaming, there’s nothing like AI to bring us together to have these conversations and dialogue and really explore, is that an okay response? Are we on side here? What are the boundaries? And there’s some really interesting research happening about building that kind of code of conduct, or what is the right mindset that you want a general-purpose AI system to have. So, we haven’t really figured that out in how we relate together as humans. It’s going to be an ambition for the AI to get there too.
Mike Hrycyk (20:21):
So, Jason, do you have an example of when you’re testing where you found that it was highlighted that human intervention was necessary?
Jason Kicknosway (20:30):
Mostly, it was working with TELUS because we were discovering as we were trucking forward. So, it was always great to receive the feedback from TELUS because when you’re wandering in the dark, you don’t know which direction you’re going. TELUS made it so easy to work on this project because they gave us really good direction like, “This is what we want you guys to look at and focus on. ” And then that helped us as we went through trying to figure out where all the biases were. Or how agentic AI works, so that’s all the little copilots that they all made. It was interesting to see how TELUS employees created things for themselves to help them out in their own daily lives. That also educated us on what we could use AI also in our daily lives.
(21:12):
When we were doing the script recording, it was very interesting because we tested in different tempos. So, we talked faster, we talked slower through the script. As you can tell, I have kind of a clinical voice when talking. So, sometimes I would be the clinician, the professional person on the other side of the phone call, and I would be the one answering all the questions, either speaking fast, speaking slower, using more country voices, more rural, and then city voices just to make sure we tested everything as this AI came through. A lot of times – there were some instances we had the AI, it would not hear certain phrases, but just because of the way we pronounced it. So, it’s always like little nuances that AI has when you’re dealing with certain things. So, everything from math to speech, you just have to remember that AI does have some nuances and limitations.
Jesslyn Dymond (22:03):
I forgot there was a voice component. That was a really unique exercise actually because a lot of language models are AI-powered, but looking at speech-to-text is a really, really big use case. And we talk about multimodal AI now, where we’re generating images and videos, and it’s just wild the variety of outcomes that we need to plan for.
Mike Hrycyk (22:25):
Who would have thought that our testers were suddenly going to have to become actors and accent professionals? Interesting. Jason, you ended up participating in a couple of learning workshops. Just tell us quickly about those.
Jason Kicknosway (22:38):
So, one was from TELUS. So, we sat down and looked at the good and bad. So, we spent a day talking with everybody at a roundtable about what would be good ethics in AI usage. And it was a very interesting discussion. It was very interesting to see how policy people looked at the situation compared to community people, compared to tech people. Our inputs were all different, but we’re kind of going in the same direction, which mostly is just making sure that the data was protected and not free-flowing and that everybody could access the data.
(23:12):
And then I went to Mila [Quebec Artificial Intelligence Institute]. So, I went to a program called Indigenous Pathways to AI. This program was about a couple of months, and it taught us everything from the programming side, how AI started, to how you implement and use AI. So, this is everything from foundational core models. So, we looked at exactly how LLMs use the core models to produce outputs.
Jesslyn Dymond (23:38):
We’re so lucky in Canada to have so many opportunities, like what Mila has put together with their Indigenous Pathways to AI. Mila is one of Canada’s three AI centers of excellence. So, as part of the Government of Canada’s national AI strategy, they made investments in Mila, in Montreal, which was led by Yoshua Bengio and the Vector Institute here in Toronto, led by Geoffrey Hinton [Co-Founder], and these are the godfathers of AI. Really fortunate to have that research and expertise here in Canada. And then out west in Edmonton, there is AMII, which is another center of excellence for AI, and they are offering so many learning courses and research opportunities. We are really, really lucky to have that cross-country work happening in AI literacy.
(24:30):
And the workshop that Jason mentioned at TELUS is part of our work to really listen and understand the opportunities in Indigenous AI across Canada, and where we need to do more work and more education, and create new capabilities that are respectful and reflective of the distinct Indigenous groups across Canada.
(24:54):
So, being intentional, working with First Nations, Inuit and Métis peoples, hearing about where there is concern, a little bit of, I think, uncertainty about how the technology may be reappropriating systems of knowledge, data sovereignty. This is extremely, extremely important to consider. And also, a lot of opportunity to take advantage of where this technology can help leapfrog some of the institutional and historical barriers and biases that have been built into the system. And with generative AI tools, you can really create accelerated processes and unlock some of that productivity that is such a need in the Indigenous economy across Canada.
(25:39):
So, balancing the different opportunities and risks is a really large part of our own data ethics and AI strategy, but something that is, I think, being discussed at many tables across Canada. And some really important research and work is happening there through institutions like the First Nations Information Governance Center, or Indigenous AI, and Concordia University has a specific research pathway along those lines as well. So, it’s really exciting to see the possibilities when we get together and really have Indigenous-led innovation in this space
Mike Hrycyk (26:12):
So, one of the main things that we teach QAs to do is to put themselves in the shoes of an end user, and that’s one of the capabilities we build. We spend a lot of time helping people do that. And so, when you talked about sitting down at the table at the TELUS workshop, Jason, one of the things that you were learning was you were learning the perspectives of the different people who might use the stuff. And that helps the QA journey of putting themselves in those shoes because if you aren’t aware that those perspectives exist, how can you try and attempt to do that?
(26:44):
But it also highlights the idea that you can only go so far as an individual with your own life experiences to put yourself in someone else’s shoes. And one of the things that I’ve learned in my 10 years at PLATO is that I am just not capable of putting myself into an Indigenous person’s shoes. I just do not have that capability. I do not have the 300 years, 10,000 years, of life experiences that have built the way an Indigenous person lives their life. And so, for TELUS to be willing to understand that those experiences exist and bring that in, I certainly applaud that.
(27:17):
But then all of this wraps into a different interesting thing. One of the things that intelligent users of AI understand is that part of your interaction with AI is instructing it what persona you’re expecting it to use. Because just leave it in general, you might not know what perspective it’s going to give you. And so, if you say you’re looking at this as a scientist, you’re looking at this as a user, that really helps AI and guides AI. And so, all of that coming together right here, right now, that’s pretty cool. I really, really like that.
Jesslyn Dymond (27:46):
Yeah. And understanding the possibilities of those personas. Another really interesting project we have underway at TELUS is a partnership with Wâsikan Kisewâtisiwin, which is a First Nations company out of Edmonton that is developing AI guardrails from an Indigenous perspective, because what we have observed – and TELUS has an Indigenous Advisory Council, which has really been pushing us to build these partnerships and to also recognize the limitations and explore the possibilities. And the challenge with AI is that it is a tool which generalizes, and that can be extremely problematic. And it’s trained on historically biased information and just will create this version of an Indigenous person that is extremely incorrect and harmful. And so, we have to be very careful about how we set up our own policies and practices and also very innovative in learning about how there is tooling that’s informed and built through ceremony with Indigenous perspectives at the forefront to help really action and correct what AI will do by default.
Mike Hrycyk (28:54):
Heaven forbid you use the media representation of Indigenous people from the 50s, 60s, 70s, 80s, 90s, and 2000s. It’s just nuts.
(29:02):
So, one question for our wrap-up question, sort of to you, Jesslyn, is when a lot of conversations happen about AI these days, they’re centred around ROI. How is this going to save me money? How is this going to save me time? And it doesn’t feel like our conversation today is focused on that, which is great, but how do you – in talking to the uppity ups, how do you translate what you’ve been doing into the types of things that they want to hear?
Jesslyn Dymond (29:29):
So, being part of a team that’s focused on privacy, compliance, data governance, records management, all of those great risk avoidance activities, this is something we are very familiar with navigating because what we’re doing in addition to complying with legal requirements is helping avoid the risk of something happening in a way that would be very unhelpful to our business, to our revenue, and to our brand, ultimately. And so, TELUS is committed to making the future friendly, and this is the work that is going to help us bring about that friendly future together. And it is about moving forward and showing the possibility of technology when it is built well.
(30:18):
There’s nothing more frustrating than seeing projects or technologies not deliver on the promises or the outcomes that have been so intentionally planned, but you have a data breach, you have a system perform in an unexpected or harmful way, and everything is lost because of that oversight. So, we go into using AI, eyes wide open, planning for the unexpected, ensuring that we have these guardrails, these safeguards in place and tested to deliver robust and sustainable outcomes that will deliver the full potential of what is possible when we get it right.
Mike Hrycyk (30:56):
Okay. So, wrap-up question. For you, Jesslyn, this is at the company level, and you’re allowed to talk about personal as well, but there’s so many challenges for companies that are just taking their baby steps towards AI. And what is the advice that you can give those people looking towards their first induction or the first steps in AI?
Jesslyn Dymond (31:14):
Well, I think you should join a Purple Team. I mean, my favourite side effect of doing this work is that it gets you using the AI and understanding its capabilities. Where it’s helpful, where there is still work to do. But AI is really about co-creation, and this can apply to each of us in our own career journeys as companies figuring out how to build products for the future. It will take that kind of development together to understand where AI can add value and find its opportunity to help just shift the way that things are delivered in hopefully far more convenient, personalized, relevant, and respectful ways. So, you’ve just got to try it.
Mike Hrycyk (32:00):
Awesome. Slight shift to the same question to you, Jason. Talking to our listeners who haven’t started down their own personal journey, what do you recommend they do to start their journey into AI? And you’re not allowed to say, “Join a Purple Team”. That’s a great option, but not everyone has it.
Jason Kicknosway (32:15):
I would say just whatever AI you have access to, just start trying to use it. You have to realize that sometimes, when you’re starting to use AI, it’ll be like interacting with a young Sheldon. It’ll be really smart, but you also have to learn how to interact with the AI. So, in a sense, you’re learning yourself how to talk to it, how to correctly prompt it to get what you want out of it. If not, it’s just like fighting with a youngster, trying to put stuff in, and what you’re getting back is hallucinations, pretty much made-up information, is what’s coming back to you. My suggestion would be just to start with it and always get the professional package that goes along with it to secure your data.
Mike Hrycyk (32:59):
So, I’d like to thank our panel for joining us for a really great discussion about AI today. I feel that a lot of talks about AI really focus on the technical, the future, and what we can do. And this discussion here really brought in the human-centric notions and how to use it to make our interactions with each other better. And I really like that tone. I really think that needs to be part of more of the discussions.
(33:18):
If you had anything you’d like to add to our conversation, we’d love to hear your feedback, comments, and questions. You can find us at PLATO Testing on LinkedIn, Facebook, and on our website. You can find links to all of our social media and website in the episode description. If anyone out there wants to join in on one of our podcast chats or has a topic that they’d like us to address, please reach out. If you are enjoying listening to our technology-focused podcast, we’d love it if you could rate and review PLATO Panel Talks on whatever platform you’re listening on. Thank you again for listening, and we’ll talk to you again next time.