Many who work in the tech industry are bullish about artificial intelligence and its future. Some have concerns.
We get the views of two people who work within Spokane's tech industry.
Clayton Gravatt is a data and AI architect at IntelliTect, which is a financial supporter of Spokane Public Radio. Dan Olvera is a software engineer there.
DN: How would you explain to folks older like me about the benefits of artificial intelligence and maybe your concerns about it?
Clayton Gravatt: Benefits of AI. As with any powerful tool, when used wisely, AI is awesome for automating the things that you don't want to have to do by hand. It's also potentially risky in that it's very, very easy and tempting to hand over your thinking to AI in a lot of places.
Navigating that line of how do I give AI more access and involve it more in my processes so that they can go faster and I can focus on more of the things that I care about, that'd be the benefit, but not so much of the control over the system that I don't know what's going on or AI has even the capability of making permanent mistakes that negatively affect people.
Dan Olvera: And I think when you talk about concerns for AI, there's kind of two major things. There's what AI can do, and then the effects of AI.
So economic effects and displacement of workers is a major concern. As technical people, we're more on the other side of it. When AI has access, what is it doing with that data? So whether it's leaking data or it's finding holes where you didn't think there were, that's kind of where we have more expertise in the concerns.
I think there are broader concerns that are more widespread. What does people's sentiment around AI lead to at an economic standpoint? But yeah, I think our concerns, and at least my expertise where we do our pros and cons is around how much data do you give it access to and what capabilities? So read and write, very different. If it can read systems, especially if it's not confidential data, there's little risk.
But as soon as you put confidential data or writing systems or being able to delete things, that's when you get to be a lot more judicious about how you expose your data. And so enterprises are grappling all over the place with, okay, what systems do we enable and flip on all these AI switches? And then there's the capabilities as well.
What are they doing within those systems?
DN: Are there places that you consciously in your own lives choose not to stay away from AI?
CG: Definitely. So in interactions with my son, so I have a one-year-old son and I love spending time with him and reading books to him. And I've seen articles on the internet of people saying, oh, I've created this app that will read my children their bedtime stories. And to me that, sure, it's automating the function that I'm providing of reading out the story, but it loses that human connection that's really the driving value behind it. I'm not reading a book to my son for him just to hear the words, I'm doing it to connect with him.
So that human level of connection is something that I am very cognizant about not replacing with AI whenever I can.
DO: I think the other major piece is like the creative aspect, right? I do a lot of writing. Obviously, you know, in school, writing was never my strongest suit, but I really liked writing.
And then in college, you know, you do lots of technical writing courses and technical writing can be, you know, a lot of people use AI. That's one of the major, it's like the top three use cases of AI within the business right now is writing. And then research is up there as well.
But then creative writing is a very different thing to me. So like, if I'm doing that on the weekend, why am I offloading something that I'm choosing to engage with for fun to AI, right? Then that removes and it results in a lot of things that like I'm not connected to as well. So anything you want to have like, kind of like an intimate connection to should be done by yourself, at least in my opinion.
DN: I love AI for turning audio. This interview after we're done is going to go into an AI generator of some kind, and it's going to create within a couple of minutes, a transcript so I don't have to spend an hour or an hour and a half transcribing it myself. What are some of the other most useful, less dangerous uses of AI that people might ask you about that you say, this is what AI is meant to be and it's perfect for?
CG: Oh man, anything where it's easy to verify the results, verify it was done correctly.
So something like a transcript is something where you heard the interview yourself and you can quickly review that transcript. I think that's a great use case for it.
So summarization, helping you understand material, helping you do research, especially when AI is giving you links to things, for example, and you are then verifying afterwards of yourself. Any of those.
DO: We do timekeeping as consultants. So I go and track my time. I have a calendar that says, here's my meetings. I kind of know what I'm doing throughout the day. And then you have to write descriptions because then that is part of what you build a client. Writing those descriptions and pulling from a calendar and looking at all my time, I have to go to four different systems to do that, especially if I'm on multiple different clients or whatever. So things that are like administrivia, I feel like a lot of times can be leveraged with AI.
DN: As you guys continue your careers and as AI has developed so rapidly, how do you keep up with what's going on in AI so that you feel like you're constantly current and can help the clients that you work with?
CG: For me, I dedicate time every day. I start my morning with making sure I keep up on the latest developments, but there are so many that I cannot keep up on everything.
So I've developed an area of expertise, a set of tools that I am most interested in, and I'm sure to read in depth about those whenever those update. And I am tangentially aware of what else is going on in the world, but I'm intentionally not becoming an expert on those other pieces that I've become disinterested in.
DO: And I think the other part of that is because it's changing so rapidly, everyone feels behind and that's almost like a feature, not a bug of the system, where because you feel behind, then you don't need to go invest and put money into it and that's just because companies are trying to one-up each other, right?
So pick your niches and pick where your strengths are and focus there on developments or new things and new features that come out. Because as things continue to go, lots of things are gonna go by the wayside, right? We're gonna find the better ways to leverage AI, right?
Ten years ago, everyone was like, oh, we're gonna use voice chat all the time with our phones. And it was terrible and everyone hated it and you looked like, oh, this is not working anymore. So then it kind of fell by the wayside. And now it's like coming back, but there's a lot of things that people are trying out, but not necessarily going to stick around for a long time.
DN: So is it your job to help companies wade through all of that and determine how quickly should I be adopting, or should I just leave this alone and wait for a while until the bugs are worked out, so to speak?
DO: Depends on the role, right? Some companies can do that, right? They're the more agile, they want to be at the front lines and experiencing things. Other companies aren't, they're a financial institution, they're not gonna be picking up the latest thing just because it looks cool, right? They're gonna move slowly, they have more regulations. And that's a good thing, right? They should be picking up the things that are tried and true. So yeah, definitely depending on the role.
I think it's our job as software people to be informed of what's going on in the technical space and how to, I think we're constantly trying to re-evaluate how we work with the tools daily. And there's obviously a pressure just externally as well. People are seeing these great exponential productivity gains. Okay, well, give it time, see if it works. If not, let's pivot and try something else.
CG: I do think that this is an important part of my job. I do spend a lot of time working with companies delivering proof of concepts, so little prototypes to see if AI actually does work. There is a lot of marketing and a lot of hype and a lot of misinformation about it. And so I think an important part before diving into fully embracing an AI thing, whatever that is that a company is, trying it first, proving that it actually has the benefits that you think it will, and then moving forward with adopting it.
DN: I move forward with AI, but I don't know that I am. It's just because everything that I use has adopted parts of AI, and I just use part of what they do. I don't even think about, oh, this is AI. Is that sort of a normal thing?
DO: Yeah, I think as companies move towards it because they need to just be able to say that we have AI, that's a shareholder thing, that's just a product of momentum within the economy. People want AI, or they want to be able to say they have AI, and so as we're interfacing as consumers, we're gonna be interacting with AI every day. Google no longer is just Google. Google always has an AI overview, so I don't think that's like, oh, am I doing something wrong, or am I going out of my way to use AI? Everyone is pretty much interfacing with AI at some point in their day.
Now, what you choose to do with that is obviously different, right? The companies are doing one thing, how you want to work. Or like, I have my own boundaries around creative writing, or time with people and personal connections, are very different choices, right?
DN: Different companies, different industries are talking about we need to have an AI policy, and that will look different for every industry or every company. What are some of the questions you would encourage them to consider as they think about what kind of policy should we have about AI usage as an organization and for our individual employees?
CG: So from an organizational perspective, one of the top concerns is often about data privacy and data usage, so making sure that whatever tools you are adopting are protecting your data and keeping it private to the level of privacy that your data requires.
For me, a lot of it is perspective and making sure that people are thinking about the tools the right way.
So it's hard to enforce this from a policy perspective. It's easier to do this through empowering individual users and having them disseminate that knowledge throughout the industry, but finding ways that are actually practical to apply AI tools.
It's hard to do that from a top-down perspective because a lot of the people in the highest levels that are saying, oh, we're gonna use this tool may be a little bit further removed from the day-to-day work, and so both having buy-in from a top-level perspective of here are tools that are safe, that meet our organization's compliance requirements, but also here are the tools that our actual workers that are going to be using these tools, so validated as, yes, this is useful for me, now let's actually adopt it, as opposed to adopting AI for AI's sake, again, proving out the value of AI for that day-to-day work as well.
DO: Yeah, statistically, majority of AI initiatives that succeed are employee-driven. They're something that, hey, I am at the ground floor, I'm doing something every day, and I've automated it from some process that's given by a tool, and then I push that up to become a policy, right?
I think there's a balance of give freedom to be able to learn tools and be able to automate and be able to innovate, but once it's then, hey, this is something that's helpful and interesting, then regulate it and make sure that it's safe and compliant, and that's where a lot of the actual engineering goes into it, right? Once you know that, well, if I have this data exposed to AI in some sort of way, it's really helpful. Then there needs to be a lot of work done on the authentication and the security side around, okay, then what kind of users are using it? What are the permissions around it? What's the data latency? All the SLAs that are actually required for enterprise data, but you first, like Clayton said, go out and prove the thing that it works.
DN: How quickly is your job changing? It used to be you could work 20, 30 years at the same job, and you didn't have to change, but it almost seems like day-to-day, your job, your industry is changing. Do you have any idea where you'll be two or three years from now? Will you be doing the same thing you're doing now? I ask that rhetorically.
CG: Yeah, I'm sure what I'm doing will change a little bit. I think that the role of a software engineer or a data person is changing, but I think that the concept of the role, at least, will still exist.
A vast majority of the jobs, I think it's something like 60% of jobs that people hold today did not exist 50 years ago. So even though it feels like technology change will lead to something like job loss, in a lot of cases, it just leads to change in jobs. And so for our jobs, I don't know about your opinion on this, Dan, but my job has changed significantly in terms of how I'm allocating my time throughout the day and the mechanical tasks that I'm expected to complete.
Specifically, since about November of 2025, there were some specific jumps in agentic coding capabilities that happened around that time to where it became more worthwhile to have agents write all of the code and then you review it rather than you write some of the code.
So I think that's probably the largest change that has happened since, is I and the majority of other developers that I know are now spending more time reviewing things than actually going through a heavy keystroke of writing it out.
Now, there's a lot of the job that hasn't changed though. There's still the problems of figuring out what should be built at all, figuring out how the larger scale pieces should fit together, figuring out how to manage risk, how to test the system and I don't know that a lot of those backing ideas will go away as AI changes. They could, but I don't think so. I think it's more so that those individual mechanical tasks change. So the act of writing code has definitely taken up less of my day. It would be something like four to six hours of my day before and now it's more like 15 minutes maybe of typing.
CG: I want to touch on that very common concerns. Is my job going away? And I like to draw on this idea called the doorman fallacy, which is a common mistake of thinking that the value of a whole role is just its most obvious mechanical tasks. So the obvious example is a doorman. Well, what do they do? They open the door. So we can just replace them with an automated door, right? And we'll be more efficient because the door never gets tired and it's cheaper to operate.
But the reality is that there's a lot of hidden value that those doormen provide too. They help greet customers. They are a level of security, just their presence. They are a status symbol.
And I think this is true for a lot of jobs as well that there is, of course, an obvious mechanical portion of it, like writing code for software engineers, but there's a lot under the surface that won't be replaced by automation.
DO: Yeah, I think definitely that people's roles are more than what is visible. I do think there is potential for a displacement period, right? Where, because they don't, or higher-upms don't see that, they only see the visible value, that I think people, a lot of people need to be up-skilled.
DN: Okay. So in the last couple of minutes, tangentially, it seems to me that we are, as a society, more aware of how AI is changing us. We still have to have human beings.
DO: if you have a fully automated system, I would guess you're going to be outperformed by those that make use of engineers and really smart people to make those decisions instead, right?
AI is based on training data, and the training data, just by law of averages, will result in the most mediocre, most, like, middle-of-the-road solution, right? If you need innovation, I am not someone that thinks that can be devoid of human thought and creativity. So it's not necessarily that we're fighting for our place and jockeying for a place in the equation. It's more that the value is there, and it's more about understanding and making that visible.
CG: I do think, though, that it's a bit of a trap to think that AI will always be the most average and most mediocre solution. So more modern means of training AI relies on something called reinforcement learning, where instead of telling AI an answer, you are instead giving AI some objective, and AI has to learn how to get to that objective. And in that training paradigm, it is possible for AI to think, or not think, sorry, to happen upon and find more efficient paths than humans may have provided. So I don't think we're safe to think that AI will always be the most average. I think that there are cases where AI may become, perhaps, superhuman at some things, and in some cases, it already is. Like, it's a lot faster, definitely.
DO: Oh, absolutely, yeah.
CG: Some niche cases, I think it is better than humans at some understanding of, like, parsing through a thousand documents and finding interesting patterns within. However, even if AI is amazing at something, there are other constraints beyond just the technical capability. An example might be something like commercial airline pilots, that on airplanes, a lot of the flight is done by autopilots and even if autopilot got to the point where it was better at takeoff and landing than pilots were, I don't think that any passenger is gonna get on a plane that doesn't have a pilot behind the wheel. And so I think that there are some more cultural and perhaps policy-driven safeguards that can exist that will help keep people in the loop beyond just us being more technically capable than robots at a given time.
DO: Yeah, and I think that's more what I was, it's not that we need to say, here's our value, and therefore it's greater than the value AI provide, it's more, as a holistic answer, we cannot just say, well, those values are the same, right? There's a level of human judgment and trust that you have behind someone that is trained and good at the thing beyond just their outcome is worse than AI outcome.
