Inside the Numbers: AI Risk in Engineering with Jim Compton and Joe Bates

WEBVTT
00:00:05.393 --> 00:00:08.812
Welcome to the Engineering Influence Podcast from the American Council of Engineering
00:00:08.812 --> 00:00:13.332
Companies. And today we're digging into a new study from the ACEC Research Institute
00:00:13.332 --> 00:00:16.614
called Managing AI Risk in Engineering Firms.
00:00:17.090 --> 00:00:20.912
Research built on interviews with firm leaders, public infrastructure owners,
00:00:20.912 --> 00:00:24.689
technology vendors, insurers, attorneys, and licensing authorities.
00:00:25.078 --> 00:00:28.612
All trying to answer the same question, which is what actually happens when
00:00:28.612 --> 00:00:32.223
AI enters a profession where the standard of care isn't optional.
00:00:32.793 --> 00:00:37.412
To help us unpack that, I'm joined by two guests. If you're any fan or follower
00:00:37.412 --> 00:00:39.768
of the program, the man needs no introduction.
00:00:40.053 --> 00:00:43.785
Joe Bates, our senior researcher at the ACEC Research Institute,
00:00:44.087 --> 00:00:48.251
a longtime returning guest, and has been instrumental in the Institute's work
00:00:48.547 --> 00:00:52.110
on the AI series of research they've been doing, the Firm of the Future.
00:00:52.893 --> 00:00:57.578
Larger initiative with the Institute and all of its economic research.
00:00:57.793 --> 00:01:03.612
Joe has been guiding the industry research for the Institute for a number of
00:01:03.612 --> 00:01:07.032
years now, And it was great to have him back on the program.
00:01:07.032 --> 00:01:09.995
And joining us with Joe is Jim Compton.
00:01:10.599 --> 00:01:15.302
He's the Chief Technology Officer at ESP Associates. That's a multidisciplinary
00:01:15.302 --> 00:01:17.222
engineering, surveying, and planning firm.
00:01:17.635 --> 00:01:21.132
Jim's a licensed professional engineer with more than 35 years of experience
00:01:21.132 --> 00:01:23.643
across engineering, IT, and business development.
00:01:24.183 --> 00:01:29.052
And he leads ESP's technology strategy, IT infrastructure, and cybersecurity
00:01:29.052 --> 00:01:34.202
initiatives, which makes him a great person to talk about what AI governance
00:01:34.202 --> 00:01:37.182
actually looks like inside of a working engineering firm.
00:01:37.182 --> 00:01:41.968
So Jim and Joe, welcome both of you to the show. Thanks, Jim.
00:01:42.872 --> 00:01:45.754
Well, let's go back to the start here.
00:01:46.748 --> 00:01:52.782
Joe, if you can kind of give us a kernel of where this research came from, the specific focus on,
00:01:53.867 --> 00:01:58.669
really taking it away from just the technology and our firm's adopting technology
00:01:58.669 --> 00:02:02.469
to the actual implementation and governance within firms.
00:02:02.469 --> 00:02:08.065
And what risks did you expect engineering firms to be most concerned about?
00:02:09.229 --> 00:02:13.540
And was that validated through a lot of the interviews that you did?
00:02:14.309 --> 00:02:23.009
Yeah. So we started out really trying to identify what the risks of utilizing
00:02:23.009 --> 00:02:28.256
AI are across the firm from a variety of different perspectives.
00:02:28.796 --> 00:02:33.269
And we expected at the beginning that we might come out with a list to sort
00:02:33.269 --> 00:02:36.524
of document various risks that are out there.
00:02:37.269 --> 00:02:43.709
As we progress through the research, we realize if we identify risks today while
00:02:43.709 --> 00:02:49.562
we are conducting the research, those could be different six months down the road, right?
00:02:49.898 --> 00:02:57.919
So we took a different approach and we created a framework for how firms can
00:02:57.919 --> 00:03:02.466
identify the risks. And we do identify some risks within the research,
00:03:02.663 --> 00:03:03.909
but more importantly, we're going,
00:03:04.533 --> 00:03:11.946
How can firms take an organizational approach to managing the risks associated with AI?
00:03:11.946 --> 00:03:18.416
Because we also learned very, very quickly in the research that AI is not just a technology issue.
00:03:18.416 --> 00:03:22.606
You could have a lot of different AI programs, software agents,
00:03:22.606 --> 00:03:27.677
and if you don't have a governance structure, if you don't have a strategy,
00:03:27.979 --> 00:03:31.986
then you're not going to be able to effectively use AI properly.
00:03:31.986 --> 00:03:35.336
And you're not going to be able to effectively address the risks that come and
00:03:35.336 --> 00:03:39.258
go as we progress through the mature life cycle. Yeah.
00:03:40.367 --> 00:03:45.476
And I want Jim to come in on this because both of your experience in technology,
00:03:46.196 --> 00:03:49.576
and as a firm leader, but starting with governance.
00:03:49.576 --> 00:03:54.896
You know, if a firm is using AI today, and I think we can all assume that if
00:03:54.896 --> 00:04:00.426
you have a firm that has a connection to the Internet, someone's using AI in some form or fashion.
00:04:00.426 --> 00:04:05.644
But let's say a firm is using AI but does not yet have a mature AI governance structure.
00:04:06.225 --> 00:04:11.476
What specifically should it put in place first to reduce the initial risk and
00:04:11.476 --> 00:04:16.716
kind of tourniquet first, you know, triage before it starts building out that
00:04:16.716 --> 00:04:20.086
larger, mature structure? What would some of those controls,
00:04:20.086 --> 00:04:21.192
what would they look like?
00:04:30.051 --> 00:04:34.476
The first thing that we do i don't think people should do is name a person to
00:04:34.476 --> 00:04:39.276
be in charge i mean you need someone to own this and it can't be a group it
00:04:39.276 --> 00:04:42.169
can't be a committee it needs to be somebody,
00:04:42.848 --> 00:04:48.716
a specific person that's actually knowledgeable about ai they're excited about.
00:04:50.029 --> 00:04:53.866
They genuinely care about the growth of the company, but they also care about limiting risk.
00:04:54.266 --> 00:04:58.759
And so you can't just pick whoever's available. So once you've got someone in charge...
00:04:59.610 --> 00:05:03.171
You need to create a small cross-functional group of some engineers,
00:05:03.171 --> 00:05:09.136
some IT, some legal people, and put this AI owner in charge of that group.
00:05:09.659 --> 00:05:14.643
And then they're going to be able to write a policy that addresses all the different
00:05:15.032 --> 00:05:18.971
issues in the company. But the policy doesn't have to be something huge to begin
00:05:18.971 --> 00:05:23.618
with. I mean, it could be short. I mean, you just need an acceptable use policy.
00:05:24.338 --> 00:05:27.281
You can start out with one or two pages. It doesn't need to be 50 pages.
00:05:27.281 --> 00:05:28.314
No one's going to read it.
00:05:28.668 --> 00:05:33.393
So you need to cover, you know, what tools are approved for use,
00:05:33.625 --> 00:05:38.001
what data can go in and what can't, who do you call or who do you reach out
00:05:38.001 --> 00:05:42.011
to if you want permission to get a license? You know, who approves a project
00:05:42.011 --> 00:05:43.628
if I've got a project I want to do?
00:05:43.750 --> 00:05:48.713
Those kind of things need to be in there so you can get some control around
00:05:48.922 --> 00:05:52.371
what's happening with AI. But before you can approve a project,
00:05:52.371 --> 00:05:55.116
I mean, you sort of need to know how much time it's going to take to create,
00:05:55.779 --> 00:05:58.170
the project and how much time it's going to save you.
00:05:58.838 --> 00:06:00.991
And initially, you're going to be guessing at this stuff. You're not going to
00:06:00.991 --> 00:06:04.061
know, but your guesses will get better over time.
00:06:04.258 --> 00:06:08.621
And, you know, is it a one-time task somebody wants to automate or is it something
00:06:08.621 --> 00:06:09.663
you don't in every project?
00:06:10.040 --> 00:06:15.971
So the projects that are available early on, the decisions should be pretty
00:06:15.971 --> 00:06:21.411
early, should be pretty easy because there's so many things that you can do with AI.
00:06:21.411 --> 00:06:24.501
But then you got to get in front of your project managers and tell them all
00:06:24.501 --> 00:06:30.593
about your policy and try to get your hands around the thing.
00:06:31.952 --> 00:06:34.332
Because at some point, someone's going to come to you and ask you,
00:06:34.999 --> 00:06:37.666
you know, you approved me to do this project. What do I charge my time to?
00:06:37.666 --> 00:06:41.739
Where do I put that? You know, how do we track this stuff that we're working on?
00:06:42.227 --> 00:06:45.596
You know, is there a project number I can charge it to? And I would recommend
00:06:45.596 --> 00:06:48.746
if someone is going to do larger projects that they do have a number to charge
00:06:48.746 --> 00:06:49.706
time to for every one of them.
00:06:49.746 --> 00:06:52.739
Because otherwise, you're never going to be able to track it.
00:06:54.274 --> 00:07:00.036
So just start with some basic stuff. And looking at and taking kind of a wider
00:07:00.036 --> 00:07:04.646
view, Joe, I mean, the study kind of looked at eight.
00:07:05.836 --> 00:07:10.696
Risk domains with AI, and it's governance, professional liability,
00:07:10.696 --> 00:07:15.328
workforce, data, the business model itself, risk to leadership.
00:07:16.263 --> 00:07:21.948
Of the eight risk domains in the framework, which do you think engineering firms,
00:07:22.563 --> 00:07:25.978
and their leaders need to most understand now, like right now,
00:07:25.978 --> 00:07:28.967
you know, as as they address this issue?
00:07:29.307 --> 00:07:33.078
It's a good question, Jeff. And that the answer is a little more complicated
00:07:33.078 --> 00:07:38.053
because all of these eight domains are connected. They're interrelated.
00:07:38.418 --> 00:07:42.662
You can't focus on one area and ignore another area.
00:07:43.324 --> 00:07:48.548
Yeah. So some of the overarching themes, though, that really come out is that
00:07:48.693 --> 00:07:52.107
you need to maintain the engineering judgment.
00:07:52.345 --> 00:07:57.790
That's probably the most important thing, right, is with AI,
00:07:58.063 --> 00:08:03.297
you can't let AI make the decisions. You can't not check AI.
00:08:03.588 --> 00:08:09.718
So you have to be able to maintain the judgment of the engineer and keep that,
00:08:10.178 --> 00:08:14.693
sort of people first is a lot of the terms that I heard when I was conducting these interviews.
00:08:15.145 --> 00:08:19.648
So it's really important to focus on that. But again, they're all interrelated.
00:08:19.648 --> 00:08:24.718
You have to also focus on workforce. How are you going to train people how to use AI?
00:08:24.979 --> 00:08:28.778
And as Jim was saying, if you have an acceptable use policy,
00:08:29.118 --> 00:08:31.759
you got to make sure everybody knows what that policy is.
00:08:31.962 --> 00:08:36.498
Then as people are learning about AI, you have to make sure there's a feedback
00:08:36.498 --> 00:08:41.338
loop, right? So that people that are experimenting, that are learning how to
00:08:41.738 --> 00:08:45.778
create tasks, they have to teach the AI to me.
00:08:46.545 --> 00:08:49.996
They have to teach the AI, but then they also have to make sure that people
00:08:49.996 --> 00:08:54.166
in the firm understand, oh, we made these improvements. We can now do this.
00:08:54.166 --> 00:08:57.330
We can save time on that. I did this and it worked really well.
00:08:57.760 --> 00:09:00.557
So there's that workforce issue and feedback loop.
00:09:00.864 --> 00:09:05.758
And as we already talked about governance, I mean, you have to have governance in place.
00:09:06.054 --> 00:09:10.976
That's that's really the first thing to do. But again, I think everything's
00:09:10.976 --> 00:09:13.838
interrelated. So you have to address it all together.
00:09:14.426 --> 00:09:19.036
Yeah. And I think that's important because it's also running a knife's edge
00:09:19.036 --> 00:09:22.806
to an extent here because you need to have a governance structure in place.
00:09:22.806 --> 00:09:27.301
You need to have policies that are clearly defined and also enforced.
00:09:27.707 --> 00:09:30.649
You can't just write it and put it on the shelf. You need to make sure that
00:09:31.032 --> 00:09:34.776
it's understood by the employee base and that it's actually being followed.
00:09:35.049 --> 00:09:40.856
But at the same time, you don't want to shut down innovation and you don't want
00:09:40.856 --> 00:09:44.686
to, you know, put a blockade up against good ideas.
00:09:44.686 --> 00:09:48.644
So, you know, Jim and first and then Joe, I mean.
00:09:49.586 --> 00:09:54.066
The risk of unmanaged AI use, you know, the shadow AI, built in tools,
00:09:54.066 --> 00:10:00.426
custom proprietary agents, workflows that kind of start as experiments,
00:10:00.426 --> 00:10:02.346
but then get into the project delivery cycle.
00:10:02.346 --> 00:10:07.299
Those all have to be guarded against, and governance does play a role in that.
00:10:07.694 --> 00:10:11.949
But now, you know, the innovation side, how do you keep that innovation cycle
00:10:12.292 --> 00:10:17.536
going when you start putting controls? What do leaders need to do so that they
00:10:17.536 --> 00:10:20.526
can still stay on the leading edge, but at the same time, they're protecting
00:10:20.526 --> 00:10:22.473
their business and their standard of care?
00:10:23.323 --> 00:10:26.660
Yeah. I mean, I want to back up a little bit there, too. I mean,
00:10:26.660 --> 00:10:29.950
you were talking about being able to enforce your procedures.
00:10:29.950 --> 00:10:34.080
I mean, you need to edit your QAQC procedures to talk about what has to have
00:10:34.080 --> 00:10:37.880
a sign-off and what has to have a review. It doesn't matter if it was AI created or not.
00:10:38.260 --> 00:10:41.360
And you have to draw a hard line there on those activities. And you have to
00:10:41.360 --> 00:10:42.420
hold people accountable.
00:10:43.052 --> 00:10:46.977
I mean, there's got to be consequences if people are not following the review,
00:10:47.476 --> 00:10:51.120
procedures because your company's reputation is on the line.
00:10:51.120 --> 00:10:52.659
I mean, you've got to do that.
00:10:53.347 --> 00:10:56.530
So, I mean, the liability isn't in AI making a mistake.
00:10:56.530 --> 00:11:01.400
It's in somebody not reviewing it and catching a mistake. But the shadow AI
00:11:01.400 --> 00:11:05.120
is a huge problem. I mean, we didn't think it was that big of a problem.
00:11:05.120 --> 00:11:09.402
But, I mean, it's basically people using public tools.
00:11:10.993 --> 00:11:13.730
But mainly because you haven't given them an approved alternative.
00:11:13.730 --> 00:11:17.660
You haven't told them, here's the company version. And so they've got a chat
00:11:17.660 --> 00:11:20.485
or a cloud account, and they're just out there doing things.
00:11:20.984 --> 00:11:23.028
You don't know what they're doing. You don't know what they're uploading.
00:11:23.242 --> 00:11:26.390
And so when we started working with AI a couple of years ago,
00:11:26.390 --> 00:11:29.500
we were wondering how much AI traffic was on our network. And so we started
00:11:29.500 --> 00:11:33.666
looking, running some tools to try to find out.
00:11:34.357 --> 00:11:40.750
And we had eight, this is several years ago, we had 80,000 AI domain hits to
00:11:40.750 --> 00:11:45.282
720 different AI tools in a four-week period.
00:11:45.914 --> 00:11:49.530
And, you know, we didn't have any approved tools. People shouldn't have been
00:11:49.530 --> 00:11:53.436
doing anything, right, according to the management, because nothing was approved.
00:11:53.779 --> 00:11:59.760
So a few concrete things you can do is buy the enterprise version of the AI
00:11:59.760 --> 00:12:04.024
tools, not the consumer versions. They have access controls in them.
00:12:04.523 --> 00:12:08.250
They let you manage who can buy tokens and who can't, how many tokens they can
00:12:08.250 --> 00:12:11.333
buy, if you know how many accounts you have.
00:12:11.910 --> 00:12:16.350
And if you don't, there's a lot of horror stories about people spending tokens,
00:12:16.350 --> 00:12:19.350
people spending $10,000, $20,000, $30,000 a month in tokens.
00:12:19.350 --> 00:12:22.961
But even the basic subscriptions are $20 a month.
00:12:23.426 --> 00:12:26.770
If people are out there putting them on their credit cards and 2,000 people
00:12:26.770 --> 00:12:28.800
do that, that's $40,000 a month.
00:12:28.800 --> 00:12:33.720
You have no idea where it's going or how to account for it. So we set up an
00:12:33.720 --> 00:12:39.767
AI governance team, and the AI owner is in charge of that team.
00:12:40.145 --> 00:12:45.270
And that team, we have an AI collaboration site to encourage people to use AI
00:12:45.270 --> 00:12:49.205
and do things with AI. They run that site. They talk with managers. They talk with groups.
00:12:49.727 --> 00:12:54.087
So they're out there interacting with people, encouraging them to use it.
00:12:54.412 --> 00:12:58.365
If somebody builds something complicated, they're trying to make sure it gets documented.
00:12:58.917 --> 00:13:03.462
How is it built? How it works? In case that person leaves, somebody's got to maintain it.
00:13:04.143 --> 00:13:07.289
So it's a process, but...
00:13:08.319 --> 00:13:16.475
You've got to get people around to experiment. We do that with policy and restrictions on files.
00:13:16.475 --> 00:13:20.283
We use the data exfiltration software. I can talk about more later to,
00:13:20.480 --> 00:13:24.755
you know, to keep confidential stuff from getting uploaded. But you've got to
00:13:24.755 --> 00:13:28.061
encourage people, but you've got to put guidelines in, too.
00:13:28.683 --> 00:13:32.455
Yeah. I mean, let's go into that because, you know, one of the things,
00:13:32.455 --> 00:13:35.075
we've had a lot of programs on AI, right?
00:13:35.075 --> 00:13:40.905
And from the early days when, you know, the LLMs were just hitting public and
00:13:40.905 --> 00:13:44.705
people were all talking about ChatGPT back in the days when,
00:13:44.705 --> 00:13:48.705
you know, the image generation would give you five, you know,
00:13:49.285 --> 00:13:53.062
more than five fingers a hand and you can tell what was fake and what was not.
00:13:53.764 --> 00:13:57.905
You know, we've come a long way in a short period of time. At that beginning
00:13:57.905 --> 00:14:02.785
stage, a lot of the conversation went around that question of organizational,
00:14:03.354 --> 00:14:05.425
knowledge, understanding data,
00:14:06.005 --> 00:14:10.599
understanding where really that transition from,
00:14:11.381 --> 00:14:15.265
you know, an engineering firm just being an engineering firm to being a data
00:14:15.265 --> 00:14:17.024
firm that does engineering.
00:14:18.825 --> 00:14:22.504
It's a question of how is that data stored? How is it governed?
00:14:23.223 --> 00:14:27.245
And then what controls are on those? So, let's go into that a little bit more
00:14:27.245 --> 00:14:32.235
about the importance of understanding good data hygiene, good data control and
00:14:32.235 --> 00:14:33.858
implementation within a firm.
00:14:34.525 --> 00:14:37.024
Explain that from your experience.
00:14:38.205 --> 00:14:43.405
As far as trying to get your data usable, I guess, in AI is what you're looking
00:14:43.405 --> 00:14:47.125
for, your knowledge management aspect of AI. I mean,
00:14:49.620 --> 00:14:53.545
that's a tough issue, right? I mean, everybody's got plans, you've got specs,
00:14:53.545 --> 00:14:57.115
you've got checklists, you've got lessons learned. A lot of that stuff's in
00:14:57.115 --> 00:15:02.013
paper, and you've got to get it so that it's usable by AI.
00:15:02.646 --> 00:15:08.565
And that's a boring task that nobody wants to jump into. But if you don't have
00:15:08.565 --> 00:15:11.486
that data usable, then...
00:15:13.873 --> 00:15:20.363
You're not going to get the full benefit out of AI if you don't have the data for it to learn from.
00:15:20.363 --> 00:15:22.693
I mean, you can do a lot of projects. You can do a lot of things.
00:15:22.693 --> 00:15:27.083
You can accelerate a lot of things, but you're not going to get the full benefit out of it.
00:15:27.083 --> 00:15:31.583
I mean, there's the only standard that I'm aware of out there.
00:15:31.583 --> 00:15:38.293
There's ISO or GIS 30401. I think it is. That's the knowledge management system
00:15:38.293 --> 00:15:40.081
standard, but it's really generic.
00:15:40.609 --> 00:15:42.879
I mean, it has a lot of good stuff in it, but it's very generic.
00:15:43.489 --> 00:15:47.363
There are pieces of software out there like Guru and Document 360 and stuff
00:15:47.363 --> 00:15:52.022
like that will help you put your knowledge into format.
00:15:53.173 --> 00:15:58.643
We don't use one. We haven't been using one. You know, we're in the early stages
00:15:58.643 --> 00:16:01.493
of trying to get that stuff done.
00:16:02.430 --> 00:16:05.274
I don't know that there's any one best way to do it.
00:16:06.087 --> 00:16:10.156
You know, honestly, we struggle with trying to get people to cooperate to get it in there.
00:16:10.156 --> 00:16:15.836
No one has the non-billable time to be loading old projects into the system.
00:16:15.836 --> 00:16:21.846
So that is a challenge. But the firm that gets their stuff organized and into
00:16:21.846 --> 00:16:24.846
AI so it can learn from it, I mean, they're the ones that are going to get the
00:16:24.846 --> 00:16:27.326
value. Everybody else is going to be left behind.
00:16:27.922 --> 00:16:32.076
Yeah, I agree with that 100%. I think the differentiator is going to be the
00:16:32.076 --> 00:16:38.896
firm that invest in the teams to definitely manage and control the data or the
00:16:38.896 --> 00:16:42.504
ones who are going to have a competitive advantage with their temporaries.
00:16:43.125 --> 00:16:46.176
I'd also say that, you know, one of the interesting things that, you know,
00:16:46.916 --> 00:16:52.881
getting into a conversation with a couple of firms in previous interviews is,
00:16:53.516 --> 00:16:59.766
kind of shocking compared to some other industries of how few firms have positions
00:16:59.766 --> 00:17:03.266
like chief information officer or chief data security officer.
00:17:03.835 --> 00:17:07.408
And the need for the industry to kind of.
00:17:08.716 --> 00:17:14.417
You know, take pace with other industries and have those positions in the C-suite
00:17:14.812 --> 00:17:19.626
that are in those executive leadership positions and have the impact that are
00:17:19.626 --> 00:17:23.114
empowered to take control of that data.
00:17:23.725 --> 00:17:25.781
You know, what's your what's your view on that, Jim?
00:17:27.551 --> 00:17:31.021
I agree. I mean, you've got to have someone. I mean, the data governance is
00:17:31.021 --> 00:17:34.871
a corporate decision. You know, it doesn't, that decision doesn't belong to
00:17:34.871 --> 00:17:38.566
IT or anybody. I mean, it's a decision the corporation has to make.
00:17:39.176 --> 00:17:43.651
And if the data is going to be the most valuable thing that you have in the
00:17:43.651 --> 00:17:46.595
future when working with AI, then someone needs to own it.
00:17:47.007 --> 00:17:50.601
Someone needs to be responsible for it. And someone needs to be challenged to
00:17:50.601 --> 00:17:53.537
get it in a format that can be used.
00:17:53.880 --> 00:17:57.258
But they also need a budget to do that, right? I mean, you can't just give,
00:17:57.409 --> 00:17:59.351
you know, you got a 2,000, 3,000
00:17:59.351 --> 00:18:02.872
person firm and you get one person to do that. That's a lot of effort.
00:18:03.383 --> 00:18:09.361
And, you know, as you're growing the AI tools, I mean, you have to treat it like anything else.
00:18:09.361 --> 00:18:13.701
You've got to pilot a test on it, see how it works, review it,
00:18:14.032 --> 00:18:18.786
and then scale it up. I mean, you don't want to go in and try to digitize all of your data all at once.
00:18:19.436 --> 00:18:22.711
And, you know, I was talking to someone recently, and they were talking about,
00:18:22.711 --> 00:18:26.331
well, you know, we think we can double the speed of this process that we're
00:18:26.331 --> 00:18:29.261
doing. That would be fantastic. If you do that on every project and you can
00:18:29.261 --> 00:18:31.707
double the speed of it, that's an amazing improvement.
00:18:32.084 --> 00:18:35.184
How do you build that project? They're like, oh, we build this all hourly.
00:18:35.803 --> 00:18:39.791
I'm like, well, I don't know you've done yourself any favor if you're doing
00:18:39.791 --> 00:18:41.636
it half the time and you're building it hourly.
00:18:42.199 --> 00:18:45.881
You know, so the first projects maybe you need to start looking at are things
00:18:45.881 --> 00:18:51.702
you build lumps on, not half in your income from doubling the speed of doing an hourly job.
00:18:52.328 --> 00:18:57.571
So it's, yeah, somebody needs to own it and make those decisions.
00:18:57.571 --> 00:19:04.077
And that person needs to be the AI owner in conjunction with someone who's in charge of the data.
00:19:05.227 --> 00:19:09.365
And I, you know, we could have an extended conversation just on this topic.
00:19:09.365 --> 00:19:13.215
It's another reason, read the study and go into it and do, you know,
00:19:13.215 --> 00:19:16.495
follow up interviews. I want to hear from the audience too. Like,
00:19:16.495 --> 00:19:18.958
what, what do you want to dig into? Because there's a lot here.
00:19:19.393 --> 00:19:24.975
I want to take a, a, a kind of zoom out a bit, Joe, and, and talk about something,
00:19:24.975 --> 00:19:27.726
which is, you know, I always kind of laugh at,
00:19:28.323 --> 00:19:32.495
to an extent, I mean, you know, you always see the futurists and the ai guys
00:19:32.495 --> 00:19:36.796
out and the west coast are talking about how work is going to become,
00:19:37.283 --> 00:19:41.525
irrelevant in a couple of years and we're all going to be replaced but the fact
00:19:41.525 --> 00:19:45.991
of the matter is i'm married to a to a lawyer who chafes at the idea as well
00:19:46.397 --> 00:19:48.591
which is that ai does not replace,
00:19:49.218 --> 00:19:51.384
the the the,
00:19:52.126 --> 00:19:59.419
the need for a professional like an engineer to be working on a project it does not,
00:20:00.632 --> 00:20:06.832
essentially lower the responsibility, the professional responsibility of the work.
00:20:07.625 --> 00:20:11.630
If anything, it just kind of puts a greater focus on that.
00:20:12.013 --> 00:20:16.165
Let's go into that and dispel that myth that, you know, AI is going to replace
00:20:16.545 --> 00:20:21.605
the judgment of a licensed professional trained engineer in the near future,
00:20:21.605 --> 00:20:22.827
because it just isn't going to happen.
00:20:23.532 --> 00:20:28.890
Yeah, we don't see that happening at all, Jeff, and this research points to.
00:20:30.573 --> 00:20:35.630
The opposite, in fact, that engineering judgment becomes more important,
00:20:35.868 --> 00:20:38.654
that it becomes more important for you to,
00:20:39.293 --> 00:20:46.053
ensure that the outputs that AI is giving you are accurate and then to interpret it, right?
00:20:46.373 --> 00:20:52.563
You have to interpret the information that the AI is giving you and you have
00:20:52.563 --> 00:20:54.722
to become that consultant to your client.
00:20:54.949 --> 00:21:01.893
And so I think it actually puts the engineer more in the center in terms of
00:21:02.213 --> 00:21:05.308
providing decisions, using their judgment.
00:21:05.447 --> 00:21:10.937
And then this has implications down the road because there are tasks that will be automated.
00:21:11.232 --> 00:21:14.356
And there are things that engineers have done for a long time,
00:21:14.890 --> 00:21:18.902
sort of in a more manual way and in a more hands-on way.
00:21:19.546 --> 00:21:23.493
And that's how they got the judgment and experience that they have,
00:21:23.493 --> 00:21:25.119
right? you have your senior engineers.
00:21:25.583 --> 00:21:28.892
So the real challenge then is how do you,
00:21:29.415 --> 00:21:35.639
change the education, the professional development process of younger engineers
00:21:36.055 --> 00:21:42.173
who may no longer have to do certain tasks that engineers used to do that helped
00:21:42.173 --> 00:21:44.020
to give them that experience, right?
00:21:44.159 --> 00:21:47.233
So again, it becomes more important, not less.
00:21:47.233 --> 00:21:51.603
And we have to focus on how we're going to develop our workforce with this new
00:21:51.603 --> 00:21:54.359
tool and that's going to change the entire firm,
00:21:54.858 --> 00:21:59.333
yeah this this this is the evolution that you know from from the the physical
00:21:59.333 --> 00:22:04.313
drafting equipment and the drafting cables to CAD from CAD to.
00:22:05.013 --> 00:22:08.946
Full digital design from digital design to AI assisted design,
00:22:09.593 --> 00:22:15.513
it is the same story told you know it's the next chapter for AI it's going to
00:22:15.513 --> 00:22:19.865
have probably a larger impact But it does come down to that,
00:22:20.381 --> 00:22:24.991
trusted advisor relationship that professional engineers have with their clients.
00:22:25.700 --> 00:22:28.616
And the study I thought was interesting that the clients really don't really
00:22:28.616 --> 00:22:30.896
care about the kind of AI that's being used, right?
00:22:30.896 --> 00:22:37.226
They're kind of agnostic on AI, but they have zero qualms about saying that
00:22:37.226 --> 00:22:41.766
we have to make sure that the people who are doing the work maintain that level
00:22:41.766 --> 00:22:44.053
of professional responsibility and judgment,
00:22:44.475 --> 00:22:46.136
regardless of what technology is.
00:22:46.980 --> 00:22:51.556
Yeah, there's a standard of care that you maintain, and that will not change.
00:22:51.556 --> 00:22:56.107
As Jim said earlier, AI by itself doesn't create the liability.
00:22:56.362 --> 00:23:02.886
The liability is if you have not checked AI and have not made sure that the
00:23:03.206 --> 00:23:06.076
outputs are correct or you've used your judgment to say, OK,
00:23:06.076 --> 00:23:07.890
I like this, but I don't like that.
00:23:08.998 --> 00:23:12.536
Yeah, I want to wrap it up there. I think that, you know, I do want to do a
00:23:12.536 --> 00:23:15.586
call out for the audience listening.
00:23:15.586 --> 00:23:19.276
I mean, tell us in the comments what you want to dig into here,
00:23:19.276 --> 00:23:26.100
because there are a lot of areas to explore. There are eight areas of risk that the study addresses.
00:23:26.660 --> 00:23:29.736
Tell us what you want to dig into, and we'll do a follow-up on it.
00:23:29.736 --> 00:23:34.774
I do want to give both you, Joe, and Jim the opportunity to wrap things up.
00:23:36.566 --> 00:23:38.976
What do you think is the big takeaway? If there had to be one takeaway from
00:23:38.976 --> 00:23:40.440
the study, what do you think that would be?
00:23:41.453 --> 00:23:46.769
You know, from my perspective, it really is that the long-term success of an
00:23:46.769 --> 00:23:53.079
organization is going to be less dependent on which AI tool they use.
00:23:53.079 --> 00:23:56.929
And it's going to be more dependent on how they manage the use of AI.
00:23:57.167 --> 00:24:02.759
That's the real critical differentiator. And, you know, Jim talked about several points.
00:24:02.759 --> 00:24:06.539
If you don't have your data organized, connected,
00:24:06.939 --> 00:24:12.179
accessible, seamlessly flowing through the project phase, If you don't capture
00:24:12.179 --> 00:24:18.079
the organizational knowledge that is not the data itself, but it is that experience,
00:24:18.079 --> 00:24:21.041
that judgment, as people retire,
00:24:21.470 --> 00:24:23.699
you're going to be behind. Yeah.
00:24:25.617 --> 00:24:31.089
And just quickly on the people part of it, I mean, if you are doubling the speed
00:24:31.089 --> 00:24:33.262
of some tasks your young engineer is doing,
00:24:33.802 --> 00:24:37.259
and you had one engineer doing the review of a group of five people,
00:24:37.259 --> 00:24:40.979
and now they're doing twice as much work, you're going to need two reviewers.
00:24:40.979 --> 00:24:46.049
I mean, it's not going to eliminate jobs. And, you know, maybe that reviewer
00:24:46.049 --> 00:24:48.089
doesn't have time to do business development anymore.
00:24:48.089 --> 00:24:51.169
Maybe he's focused on, and you have to hire a business development guy.
00:24:51.169 --> 00:24:56.409
You know, the seller-doer model may change. I mean, so there's a lot of real
00:24:56.409 --> 00:25:00.509
implications of how it does speed things up, but eliminating jobs isn't one of them.
00:25:00.639 --> 00:25:06.559
I mean, you've got to get your data in order.
00:25:07.079 --> 00:25:11.879
You've got to pilot some projects and then alter your mentoring and your training
00:25:11.879 --> 00:25:13.259
based on what you learn from that.
00:25:13.259 --> 00:25:16.392
If you run some engineers through a couple of projects and you're like,
00:25:16.781 --> 00:25:21.149
they didn't learn half of what I thought they should learn, that you need to
00:25:21.149 --> 00:25:23.519
change your training in order to deal that.
00:25:24.654 --> 00:25:28.102
Basically, I mean, AI is an amazing tool. It's getting better every day.
00:25:28.752 --> 00:25:33.133
The goal is to find some way to use it to improve the quality of service for
00:25:33.133 --> 00:25:35.311
our clients. I mean, that's the whole point.
00:25:35.933 --> 00:25:40.588
If you go back and look at a project and analyze it from a business perspective,
00:25:41.047 --> 00:25:45.283
I mean, how did it actually change how fast you're getting the work done?
00:25:45.283 --> 00:25:49.993
How did it actually change your work and what you can charge for it?
00:25:50.493 --> 00:25:53.023
And those kind of meaningful things that hit the bottom line,
00:25:53.023 --> 00:25:56.306
if it doesn't do that, then I'm not sure why we're doing it.
00:25:56.741 --> 00:26:00.913
Yeah, very good points to end the conversation with, and I 100% agree.
00:26:00.913 --> 00:26:06.023
It's not going to kill jobs. It's going to create new ones, especially for maybe
00:26:06.023 --> 00:26:11.573
the non, a lot more data people, a lot more people who are computer science
00:26:11.573 --> 00:26:15.643
and are people can work with and organize data.
00:26:15.643 --> 00:26:22.273
The people who can also help back up the engineers in utilizing the technologies.
00:26:22.273 --> 00:26:26.663
But it's definitely going to allow for maybe some greater innovation,
00:26:26.663 --> 00:26:29.473
but it's not going to reduce the world wars. And I think that,
00:26:29.473 --> 00:26:31.003
you know, that's very important.
00:26:31.456 --> 00:26:34.836
So anybody out there who's looking to get into the engineering field,
00:26:35.333 --> 00:26:37.703
probably going to be more opportunities because there are going to be different
00:26:37.703 --> 00:26:39.362
kind of jobs out there to get.
00:26:40.074 --> 00:26:44.248
I will encourage everyone to go on the ACEC Research Institute's website.
00:26:44.520 --> 00:26:50.538
We have the link to the study in the show notes for managing AI risk in engineering firms.
00:26:50.880 --> 00:26:53.693
It's a great piece of research, very usable from the C-suite,
00:26:54.099 --> 00:26:57.106
and I encourage everybody to download it and let us know what you want to hear.
00:26:57.332 --> 00:26:59.813
This is a topic that we're going to be coming back to a lot.
00:27:00.184 --> 00:27:05.573
Joe's going to be with us in fall conference. We're recording this in the last
00:27:05.573 --> 00:27:09.513
week of August and October of 2026. We're going to be in Phoenix.
00:27:10.193 --> 00:27:12.723
Joe, you'll be there talking about some economic stuff, and I'm sure we'll be
00:27:12.723 --> 00:27:13.692
touching him on this issue.
00:27:15.074 --> 00:27:17.837
Jim, I really appreciate you joining us as well and your expertise.
00:27:19.061 --> 00:27:24.033
See you next time. Fantastic. And we'll look forward to extending this conversation.
00:27:24.033 --> 00:27:27.493
So, again, this has been Engineering Influence, a podcast from the American
00:27:27.493 --> 00:27:29.968
Council of Engineering Companies. And we'll see you next time.