WEBVTT
00:00:00.160 --> 00:00:01.439
Denny, thank you so much for joining us.
00:00:01.439 --> 00:00:02.720
We really appreciate you taking time.
00:00:02.720 --> 00:00:03.680
You met Tim.
00:00:03.680 --> 00:00:04.480
Glad to be here.
00:00:04.480 --> 00:00:05.200
Absolutely.
00:00:05.200 --> 00:00:09.599
Looking back, what first inspired you to dedicate your career to behavioral health?
00:00:09.599 --> 00:00:12.800
Because you've had a pretty illustrious career in behavioral health.
00:00:13.199 --> 00:00:15.119
Or notorious, depending on how you look at it.
00:00:15.119 --> 00:00:15.679
True.
00:00:15.679 --> 00:00:22.320
No, I actually I started college as a music major, and it took one quarter to find out I wasn't quite the drummer I thought I was.
00:00:22.320 --> 00:00:27.920
And I changed as one does when you're a music major, I changed to chemistry, which, believe it or not, was better.
00:00:27.920 --> 00:00:29.679
But I really didn't know what I wanted to do.
00:00:29.679 --> 00:00:39.200
And I put myself through college, and one of the ways I did that was I got a job working in a local hospital, and they placed me on the psychiatric unit as an orderly.
00:00:39.200 --> 00:00:42.640
And that was an eye-opener for me.
00:00:42.640 --> 00:00:45.759
I mean, I've never seen people behave like that.
00:00:45.759 --> 00:00:52.719
We saw everything from depression, mild depression, to full-on schizophrenia, psychosis.
00:00:52.719 --> 00:01:03.200
And I just was amazed by it and curious and switched to psychology and got a psych and pre-med major and chemistry minor, and just that's what kicked it off.
00:01:03.200 --> 00:01:04.719
So that's how I got into this.
00:01:04.879 --> 00:01:05.439
That's cool.
00:01:05.439 --> 00:01:05.840
That's cool.
00:01:05.840 --> 00:01:14.000
I think a lot of people end up in in behavioral health on accident, but yours is a farther departure from where you started for sure.
00:01:14.000 --> 00:01:15.200
Certainly was, yeah.
00:01:15.200 --> 00:01:15.599
Yeah.
00:01:15.599 --> 00:01:17.120
F from drummer to doctor.
00:01:17.120 --> 00:01:18.959
That that could be your the name of your book.
00:01:18.959 --> 00:01:20.159
Yeah, I like that.
00:01:20.159 --> 00:01:21.200
Yeah, it's not bad.
00:01:21.200 --> 00:01:21.760
It's not bad.
00:01:21.760 --> 00:01:26.000
So you've spent decades now helping shape behavioral health technology.
00:01:26.000 --> 00:01:29.920
What innovation today excites you the most in life?
00:01:30.239 --> 00:01:33.519
Well, unless you're living under a rock, AI has gotten popular.
00:01:33.519 --> 00:01:37.280
And I think that's probably the one that is excites me the most.
00:01:37.280 --> 00:01:48.560
As a recovering clinician, I think AI for clinicians is really the most exciting thing that's happening right now, particularly with ambient listening for the helping them do their jobs.
00:01:48.560 --> 00:01:59.200
In this industry, we have seen every form of technology that's been introduced in this industry, especially electronic health records, promised to help clinicians' jobs get easier.
00:01:59.200 --> 00:02:00.319
And they never did.
00:02:00.319 --> 00:02:08.080
But AI, ambient listening AI for individual group medication management, psychotherapy sessions, is doing that.
00:02:08.080 --> 00:02:10.159
I mean, the data's clear, it is helping.
00:02:10.159 --> 00:02:33.840
And so to me, seeing clinicians who were just overburdened to begin with documentation requirements, particularly in the community behavioral health side, they seeing them get relief from that and to be able to do what only they can do, which is to provide treatment and not worry about the documentation burden that they have to face is to me one of the most exciting things I've experienced in my career.
00:02:34.240 --> 00:02:34.879
Absolutely.
00:02:34.879 --> 00:02:35.520
Absolutely.
00:02:35.520 --> 00:02:39.680
And you've often said technology should strengthen care, not replace it.
00:02:39.680 --> 00:02:44.479
What do you think most organizations are still getting wrong about AI and behavioral health?
00:02:44.800 --> 00:02:47.840
Well, there's still that worry that we saw really in all sectors.
00:02:47.840 --> 00:02:50.080
Will AI take my job, for example?
00:02:50.080 --> 00:02:53.039
Still see some of that, a little less, but still some.
00:02:53.039 --> 00:02:55.360
People have ethical concerns about AI.
00:02:55.360 --> 00:03:00.080
It was largely centered around the data center problem, which it is a problem, there's no question.
00:03:00.080 --> 00:03:16.080
But the reality is that the usage of AI like ambient listening or really anything, Facebook, Chat GPT, any of those things, the usage of those things isn't what's driving the build in data centers.
00:03:16.080 --> 00:03:19.280
It's the training models that have to have all that capacity.
00:03:19.280 --> 00:03:30.879
And so when I when we talk to people who try to point to data centers as a reason not to use ambient listening AI for their psychotherapy sessions, I say that's a personal choice.
00:03:30.879 --> 00:03:32.000
And I understand that.
00:03:32.000 --> 00:03:36.159
First recognize that that will likely be symbolic more than anything.
00:03:36.159 --> 00:03:46.479
But does that also mean that you'll stop using GPS, Facebook, Amazon, all the other online tools that we use, which you have been using AI for many years now?
00:03:46.479 --> 00:03:55.280
I mean, you Amazon provides you with suggestions of what to buy because you you bought a kayak and you now might want some ores or something.
00:03:55.280 --> 00:03:59.599
And no, I know it's listening to me, so I'll probably get an ad for kayak here in a minute.
00:03:59.599 --> 00:04:00.560
But now, yeah.
00:04:00.560 --> 00:04:02.639
But that is an AI engine on the back end.
00:04:02.639 --> 00:04:11.680
And so if you're going to be ethical about this, then you should eliminate all of those from your lives, not just I am listening for AI.
00:04:12.000 --> 00:04:12.240
Yeah.
00:04:12.240 --> 00:04:13.439
I think it makes complete sense.
00:04:13.439 --> 00:04:19.600
And I think that is a it is such a common pushback, the data centers and they're using our water and our energy and all those things.
00:04:19.600 --> 00:04:21.519
And I think that's definitely a significant issue.
00:04:21.519 --> 00:04:23.519
But I've had very similar conversations with my team.
00:04:23.519 --> 00:04:24.800
It's like, yep, absolutely.
00:04:24.800 --> 00:04:26.240
It's definitely a problem.
00:04:26.240 --> 00:04:26.879
Yeah.
00:04:26.879 --> 00:04:29.040
And we can't stop using everything online.
00:04:29.040 --> 00:04:33.120
And if we can't stop ever using everything, then we better start being smart about what we are using.
00:04:33.439 --> 00:04:45.120
And I think that I've seen some good kind of movement on that with the data center problem where Microsoft, I think, came out with a five-point program where they say we will assure that we pay for all the energy we use.
00:04:45.120 --> 00:04:46.959
We will do everything responsible.
00:04:46.959 --> 00:04:48.079
They do all the right things.
00:04:48.079 --> 00:04:57.199
And we're seeing some legislation starting to be introduced that says that a data center has to replace, they have to create more electricity than they use.
00:04:57.199 --> 00:05:00.720
And so we're starting to see some action on that front.
00:05:00.720 --> 00:05:03.759
So it's not, I'm not as pessimistic as I was for a while.
00:05:03.759 --> 00:05:11.199
There actually, my son and daughter-in-law live in a location where they're looking at getting a data center nearby and they're going to protests and all those kind of things.
00:05:11.199 --> 00:05:11.759
And I get it.
00:05:11.759 --> 00:05:12.560
I don't blame them.
00:05:12.800 --> 00:05:13.680
Yeah, absolutely.
00:05:13.680 --> 00:05:14.240
Absolutely.
00:05:14.240 --> 00:05:22.240
I do think that like all things, the first phase of expansion is always the messiest and the worst for the environment and the worst for people.
00:05:22.240 --> 00:05:24.319
And mostly things level out eventually.
00:05:24.319 --> 00:05:28.160
Shifting topics a little bit into some of your leadership experience.
00:05:28.160 --> 00:05:31.759
What have you found is the hardest part of growing your team?
00:05:32.240 --> 00:05:34.639
Yeah, I was a CEO for 17 years.
00:05:34.639 --> 00:05:37.600
And in my current role, I'm a solo practitioner.
00:05:37.600 --> 00:05:43.199
But so I call upon those experiences back when I was in the leadership roles.
00:05:43.199 --> 00:05:45.360
And I think, you know, there's a lot of things about that.
00:05:45.360 --> 00:05:47.199
One is kind of changing culture.
00:05:47.199 --> 00:05:50.000
Culture is the hardest thing to change in an organization.
00:05:50.000 --> 00:06:14.879
And then I think there's that balance, particularly at the senior level, between doing leadership in a way that engages others and that you have a participatory and actually even somewhat antagonistic set of advisors, people who will give you disparate opinions about things, but at the end of the day, it's not a democracy, and somebody has to make the final call, and that is the CEO, and that's what you're getting paid for.
00:06:14.879 --> 00:06:23.680
And so, from my perspective, I always believe that there should, I should never get a decision across my desk that's a black and white decision.
00:06:23.680 --> 00:06:26.160
That should have been already made by somebody else.
00:06:26.160 --> 00:06:31.920
Every decision I ever made was shades of gray and required a judgment call of some kind.
00:06:31.920 --> 00:06:33.040
And that's the way it should be.
00:06:33.040 --> 00:06:37.519
When I inherited my organization, my predecessor was the exact opposite.
00:06:37.519 --> 00:06:39.279
He signed everything.
00:06:39.279 --> 00:06:44.800
And I was getting five dollar chits to sign off on so somebody could get lunch.
00:06:44.800 --> 00:06:46.319
We're not doing that here anymore.
00:06:46.319 --> 00:06:50.399
That's not quite what they hired me to do, is to sign off on a $5 lunch chit.
00:06:50.399 --> 00:06:56.000
And that that was a that's what that was a cultural change that required a huge amount of communication internally.
00:06:56.000 --> 00:07:00.800
We had quarterly all staff meetings where anything was fodder for conversation.
00:07:00.800 --> 00:07:05.759
And when people said we don't like that you're doing that, I said, I am open, give me alternatives.
00:07:05.759 --> 00:07:11.680
If you have them, I'm but you can't just say I don't want to do that without giving me a constructive alternative that is realistic.
00:07:11.680 --> 00:07:15.120
So there's lots that we could talk about in terms of leadership.
00:07:15.120 --> 00:07:27.600
But I think that that issue of engaging others to be part of the decision, but also being ready as a CEO, particularly, to be able to say, thank you for your input, but I'm not we're gonna go a different direction.
00:07:27.600 --> 00:07:28.879
You have to be able to do that.
00:07:28.879 --> 00:07:29.920
That's part of it.
00:07:30.240 --> 00:07:30.800
Absolutely.
00:07:30.800 --> 00:07:35.839
That is the hard part of the job of why CEOs get paid the most in most organizations.
00:07:35.839 --> 00:07:37.199
You have to make that hard call.
00:07:37.519 --> 00:07:43.839
As a consultant, I've been in organizations and it's been just incredible to me that people, the leadership that doesn't do that.
00:07:43.839 --> 00:07:52.959
It's like I used to think of organizations, particularly in the community behavioral health space, as being mom and pop shops at one end and the large sophisticated organizations.
00:07:52.959 --> 00:07:56.560
Now I think of it as the large organizations, sophisticated.
00:07:56.560 --> 00:07:59.040
And how in the heck are you still in business, folks?
00:07:59.040 --> 00:08:00.319
Because they're doing things.
00:08:00.319 --> 00:08:06.959
It's like, and largely it's the leader who doesn't want to take the leadership role and do the things that all the leaders can do.
00:08:06.959 --> 00:08:08.720
That's kind of the job of leadership.
00:08:08.720 --> 00:08:10.399
Do your job, make hard calls.
00:08:10.399 --> 00:08:12.399
So yeah, we could have a long conversation about that.
00:08:12.560 --> 00:08:12.879
Absolutely.
00:08:12.879 --> 00:08:25.199
And to go just a little farther, I think a lot of the a lot of the time where the mistake is in leadership is it's your job to make the hard call, to make the decision, and to know when you don't need to make the decision.
00:08:25.199 --> 00:08:29.040
To make the decision to trust your team to make decisions for you.
00:08:29.040 --> 00:08:31.600
I think that's probably the scariest place for most leaders.
00:08:31.600 --> 00:08:36.639
That's like the final boss to conquer, is now passing that responsibility on, but knowing you're still responsible.
00:08:36.639 --> 00:08:40.399
If your team makes a mistake, it's still on you, even though it wasn't your decision.
00:08:40.639 --> 00:08:56.320
Well, and it's interesting you mentioned that because I think that ties together what we were talking about previously, that the whole goal of augmented intelligence is to help clinicians do what only they can do and to get the other stuff off their desk, knowing that they are still responsible for it.
00:08:56.320 --> 00:09:06.159
And it's exactly the same process and leadership that if you were working for me, I have to say I have to trust that Tim can do his job and you contact me if you want some help.
00:09:06.159 --> 00:09:08.879
But but that's why I hired you to do your job.
00:09:08.879 --> 00:09:12.399
I should not know how to do your job better than you know how to do your job.
00:09:12.399 --> 00:09:14.720
I can help you, but that it shouldn't be that way.
00:09:14.720 --> 00:09:22.960
But that's I think that is one of the concerns that a lot of clinicians have about the idea that a good AI does not write the note for you.
00:09:22.960 --> 00:09:32.960
A good AI provides suggestions and ideas, insights that you may not have noticed, and you decide what goes in there, what does not go in there.
00:09:32.960 --> 00:09:43.360
But at the end of the day, it is the clinician, when they sign that clinical note, they are attesting to the fact that note is accurate and it represents what happened in that session.
00:09:43.360 --> 00:09:52.960
And that that I think a lot of clinicians, until they see how AIs should work, they tend to think, okay, this thing's sort of write my note for me and I'm not comfortable with that.
00:09:52.960 --> 00:09:54.399
No, it doesn't write your note for you.
00:09:54.399 --> 00:09:58.080
It gives you help so you can write the note faster, easier, better.
00:09:58.080 --> 00:09:59.759
But your point is exactly right.
00:09:59.759 --> 00:10:08.159
It's like the whole goal of this is to help people work at the top end of their license or at their jobs and get the detritus off of their desks.
00:10:08.399 --> 00:10:08.879
Absolutely.
00:10:08.879 --> 00:10:09.440
Exactly.
00:10:09.440 --> 00:10:15.919
So shifting topics to the last time here, since we're a marketing agency that works in this space, we always ask a couple of marketing type questions.
00:10:15.919 --> 00:10:19.679
And because you've had so many roles, I'm sure you have some good perspectives on it.
00:10:19.679 --> 00:10:25.200
But what have you found works best when it comes to acquiring talent, especially outside of word of mouth?
00:10:25.200 --> 00:10:28.559
What works best to attract good talent to an organization?
00:10:28.960 --> 00:10:37.039
Generically, I think if you create an environment that is attractive and that you clearly are helping people do their jobs better.
00:10:37.039 --> 00:10:41.919
I always thought my job was to help people do their jobs and to get barriers out of their way.
00:10:41.919 --> 00:10:46.000
It wasn't to fix things, but it was just to remove barriers they could not remove.
00:10:46.000 --> 00:10:48.639
But I think the culture has a lot to do with it.
00:10:48.639 --> 00:10:54.480
I mean, we meant word of mouth is still one of the primary ways in which people engage in organizations.
00:10:54.480 --> 00:11:07.279
On a more specific level, what we're seeing now is that people are using, in the case of Ilios, for example, where that I consult to, organizations are now using the fact that they're using Elios as a differentiator when they recruit.
00:11:07.279 --> 00:11:12.559
So they can go out and say, you know how hard it is to write your notes at your current organization?
00:11:12.559 --> 00:11:15.120
We've got a tool that can help you expedite that.
00:11:15.120 --> 00:11:21.279
And because there's a networking among clinicians, somebody will call a friend who works there and say, is that true?
00:11:21.279 --> 00:11:22.879
Is that hype or is that really true?
00:11:22.879 --> 00:11:24.080
And no, that's true.
00:11:24.080 --> 00:11:34.559
And it really does help literally to the point where we have heard on many occasions clinicians who have told their bosses, if you get rid of this AI tool, I'm leaving.
00:11:34.559 --> 00:11:37.679
Because it's helped them that much in their in their work.
00:11:37.679 --> 00:11:41.360
So people are using it as a differentiator in recruitment right now.
00:11:41.360 --> 00:11:42.080
I love that.
00:11:42.240 --> 00:11:42.639
I love it.
00:11:42.639 --> 00:11:44.639
That's a strong case for the service.
00:11:44.639 --> 00:11:54.799
Last question then is what is one marketing tactic or strategy that you or your team is either watching or just considering in the marketplace right now?
00:11:55.120 --> 00:11:59.200
Well, I don't know that we're there's any one that we're looking at to implement.
00:11:59.200 --> 00:12:02.000
There's one particular one, though, that's really got me worried.
00:12:02.000 --> 00:12:09.519
And that is, like every technology, you mentioned this, once it gets going, there's buzz and then there's knockoffs, and there's all kinds of things.
00:12:09.519 --> 00:12:36.639
And one of the things I'm one of the things I'm really concerned about is organizations, vendors that are bringing a generic AI like ChatGPT or something off the shelf that you and I would use for our kind of day-to-day work, putting a psychological wrapper around that and selling it as a psychology-related AI, when in fact it's nothing but ChatGPT or Claude or Perplexy or something on the back end, and they just put some kind of wrapper around it.
00:12:36.639 --> 00:12:54.080
That to me is worrisome because we know that, for example, ChatGTV chat was their large language model was trained on something between 16 and 30 percent of vetted content, books, magazines, and things like that, which many of them are being sued for copyright infringement, but that's another story.
00:12:54.080 --> 00:12:56.240
But the rest of it was just scraped from the internet.
00:12:56.240 --> 00:13:02.399
And that's why you and I can log into one of those off-the-shelf tools and ask it a general question.
00:13:02.399 --> 00:13:08.720
And it's very good because it's pulled information, but it's not psychotherapy, it's not what it's not what it's created for.
00:13:08.720 --> 00:13:12.480
It's a little bit like asking Siri to interpret an x-ray.
00:13:12.480 --> 00:13:15.360
Siri does a lot of fun things and it can be very helpful.
00:13:15.360 --> 00:13:16.399
That ain't one of them.
00:13:16.399 --> 00:13:23.519
And it's the same thing here where, you know, people they conflate just general conversation and interactions.
00:13:23.519 --> 00:13:25.279
That's the same as psychotherapy.
00:13:25.279 --> 00:13:30.639
And, you know, in AIs, their base models are trained to be sycophantic.
00:13:30.639 --> 00:13:33.919
They're trained to keep you engaged and to tell you what you want to hear.
00:13:33.919 --> 00:13:42.480
Well, that's caused a lot of problems, particularly for kids and adolescents who are using these off-the-shelf tools as a quasi-therapist.
00:13:42.480 --> 00:13:59.679
And there are many cases of kids who've committed suicide in concert while they're in conversation with these AIs, and they do nothing to refer them to 988 or to get them involved in some kind of crisis prevention or to dissuade them from thinking about committing suicide.
00:13:59.679 --> 00:14:07.679
So those are probably the biggest things that I see the idea of using AI in general in a way that it wasn't meant to be used for.
00:14:08.000 --> 00:14:08.559
Absolutely.
00:14:08.559 --> 00:14:09.600
I appreciate that.
00:14:09.600 --> 00:14:10.799
I appreciate your insights.
00:14:10.799 --> 00:14:13.360
I appreciate you taking time out of your day for us.
00:14:13.360 --> 00:14:16.639
Thank you, and thank you for the work that you're doing in your community.
00:14:16.879 --> 00:14:17.360
You bet.
00:14:17.360 --> 00:14:18.080
Thanks, Tim.
00:14:18.080 --> 00:14:18.799
Glad to be here.
00:00:00.160 --> 00:00:01.439
Denny, thank you so much for joining us.
00:00:01.439 --> 00:00:02.720
We really appreciate you taking time.
00:00:02.720 --> 00:00:03.680
You met Tim.
00:00:03.680 --> 00:00:04.480
Glad to be here.
00:00:04.480 --> 00:00:05.200
Absolutely.
00:00:05.200 --> 00:00:09.599
Looking back, what first inspired you to dedicate your career to behavioral health?
00:00:09.599 --> 00:00:12.800
Because you've had a pretty illustrious career in behavioral health.
00:00:13.199 --> 00:00:15.119
Or notorious, depending on how you look at it.
00:00:15.119 --> 00:00:15.679
True.
00:00:15.679 --> 00:00:22.320
No, I actually I started college as a music major, and it took one quarter to find out I wasn't quite the drummer I thought I was.
00:00:22.320 --> 00:00:27.920
And I changed as one does when you're a music major, I changed to chemistry, which, believe it or not, was better.
00:00:27.920 --> 00:00:29.679
But I really didn't know what I wanted to do.
00:00:29.679 --> 00:00:39.200
And I put myself through college, and one of the ways I did that was I got a job working in a local hospital, and they placed me on the psychiatric unit as an orderly.
00:00:39.200 --> 00:00:42.640
And that was an eye-opener for me.
00:00:42.640 --> 00:00:45.759
I mean, I've never seen people behave like that.
00:00:45.759 --> 00:00:52.719
We saw everything from depression, mild depression, to full-on schizophrenia, psychosis.
00:00:52.719 --> 00:01:03.200
And I just was amazed by it and curious and switched to psychology and got a psych and pre-med major and chemistry minor, and just that's what kicked it off.
00:01:03.200 --> 00:01:04.719
So that's how I got into this.
00:01:04.879 --> 00:01:05.439
That's cool.
00:01:05.439 --> 00:01:05.840
That's cool.
00:01:05.840 --> 00:01:14.000
I think a lot of people end up in in behavioral health on accident, but yours is a farther departure from where you started for sure.
00:01:14.000 --> 00:01:15.200
Certainly was, yeah.
00:01:15.200 --> 00:01:15.599
Yeah.
00:01:15.599 --> 00:01:17.120
F from drummer to doctor.
00:01:17.120 --> 00:01:18.959
That that could be your the name of your book.
00:01:18.959 --> 00:01:20.159
Yeah, I like that.
00:01:20.159 --> 00:01:21.200
Yeah, it's not bad.
00:01:21.200 --> 00:01:21.760
It's not bad.
00:01:21.760 --> 00:01:26.000
So you've spent decades now helping shape behavioral health technology.
00:01:26.000 --> 00:01:29.920
What innovation today excites you the most in life?
00:01:30.239 --> 00:01:33.519
Well, unless you're living under a rock, AI has gotten popular.
00:01:33.519 --> 00:01:37.280
And I think that's probably the one that is excites me the most.
00:01:37.280 --> 00:01:48.560
As a recovering clinician, I think AI for clinicians is really the most exciting thing that's happening right now, particularly with ambient listening for the helping them do their jobs.
00:01:48.560 --> 00:01:59.200
In this industry, we have seen every form of technology that's been introduced in this industry, especially electronic health records, promised to help clinicians' jobs get easier.
00:01:59.200 --> 00:02:00.319
And they never did.
00:02:00.319 --> 00:02:08.080
But AI, ambient listening AI for individual group medication management, psychotherapy sessions, is doing that.
00:02:08.080 --> 00:02:10.159
I mean, the data's clear, it is helping.
00:02:10.159 --> 00:02:33.840
And so to me, seeing clinicians who were just overburdened to begin with documentation requirements, particularly in the community behavioral health side, they seeing them get relief from that and to be able to do what only they can do, which is to provide treatment and not worry about the documentation burden that they have to face is to me one of the most exciting things I've experienced in my career.
00:02:34.240 --> 00:02:34.879
Absolutely.
00:02:34.879 --> 00:02:35.520
Absolutely.
00:02:35.520 --> 00:02:39.680
And you've often said technology should strengthen care, not replace it.
00:02:39.680 --> 00:02:44.479
What do you think most organizations are still getting wrong about AI and behavioral health?
00:02:44.800 --> 00:02:47.840
Well, there's still that worry that we saw really in all sectors.
00:02:47.840 --> 00:02:50.080
Will AI take my job, for example?
00:02:50.080 --> 00:02:53.039
Still see some of that, a little less, but still some.
00:02:53.039 --> 00:02:55.360
People have ethical concerns about AI.
00:02:55.360 --> 00:03:00.080
It was largely centered around the data center problem, which it is a problem, there's no question.
00:03:00.080 --> 00:03:16.080
But the reality is that the usage of AI like ambient listening or really anything, Facebook, Chat GPT, any of those things, the usage of those things isn't what's driving the build in data centers.
00:03:16.080 --> 00:03:19.280
It's the training models that have to have all that capacity.
00:03:19.280 --> 00:03:30.879
And so when I when we talk to people who try to point to data centers as a reason not to use ambient listening AI for their psychotherapy sessions, I say that's a personal choice.
00:03:30.879 --> 00:03:32.000
And I understand that.
00:03:32.000 --> 00:03:36.159
First recognize that that will likely be symbolic more than anything.
00:03:36.159 --> 00:03:46.479
But does that also mean that you'll stop using GPS, Facebook, Amazon, all the other online tools that we use, which you have been using AI for many years now?
00:03:46.479 --> 00:03:55.280
I mean, you Amazon provides you with suggestions of what to buy because you you bought a kayak and you now might want some ores or something.
00:03:55.280 --> 00:03:59.599
And no, I know it's listening to me, so I'll probably get an ad for kayak here in a minute.
00:03:59.599 --> 00:04:00.560
But now, yeah.
00:04:00.560 --> 00:04:02.639
But that is an AI engine on the back end.
00:04:02.639 --> 00:04:11.680
And so if you're going to be ethical about this, then you should eliminate all of those from your lives, not just I am listening for AI.
00:04:12.000 --> 00:04:12.240
Yeah.
00:04:12.240 --> 00:04:13.439
I think it makes complete sense.
00:04:13.439 --> 00:04:19.600
And I think that is a it is such a common pushback, the data centers and they're using our water and our energy and all those things.
00:04:19.600 --> 00:04:21.519
And I think that's definitely a significant issue.
00:04:21.519 --> 00:04:23.519
But I've had very similar conversations with my team.
00:04:23.519 --> 00:04:24.800
It's like, yep, absolutely.
00:04:24.800 --> 00:04:26.240
It's definitely a problem.
00:04:26.240 --> 00:04:26.879
Yeah.
00:04:26.879 --> 00:04:29.040
And we can't stop using everything online.
00:04:29.040 --> 00:04:33.120
And if we can't stop ever using everything, then we better start being smart about what we are using.
00:04:33.439 --> 00:04:45.120
And I think that I've seen some good kind of movement on that with the data center problem where Microsoft, I think, came out with a five-point program where they say we will assure that we pay for all the energy we use.
00:04:45.120 --> 00:04:46.959
We will do everything responsible.
00:04:46.959 --> 00:04:48.079
They do all the right things.
00:04:48.079 --> 00:04:57.199
And we're seeing some legislation starting to be introduced that says that a data center has to replace, they have to create more electricity than they use.
00:04:57.199 --> 00:05:00.720
And so we're starting to see some action on that front.
00:05:00.720 --> 00:05:03.759
So it's not, I'm not as pessimistic as I was for a while.
00:05:03.759 --> 00:05:11.199
There actually, my son and daughter-in-law live in a location where they're looking at getting a data center nearby and they're going to protests and all those kind of things.
00:05:11.199 --> 00:05:11.759
And I get it.
00:05:11.759 --> 00:05:12.560
I don't blame them.
00:05:12.800 --> 00:05:13.680
Yeah, absolutely.
00:05:13.680 --> 00:05:14.240
Absolutely.
00:05:14.240 --> 00:05:22.240
I do think that like all things, the first phase of expansion is always the messiest and the worst for the environment and the worst for people.
00:05:22.240 --> 00:05:24.319
And mostly things level out eventually.
00:05:24.319 --> 00:05:28.160
Shifting topics a little bit into some of your leadership experience.
00:05:28.160 --> 00:05:31.759
What have you found is the hardest part of growing your team?
00:05:32.240 --> 00:05:34.639
Yeah, I was a CEO for 17 years.
00:05:34.639 --> 00:05:37.600
And in my current role, I'm a solo practitioner.
00:05:37.600 --> 00:05:43.199
But so I call upon those experiences back when I was in the leadership roles.
00:05:43.199 --> 00:05:45.360
And I think, you know, there's a lot of things about that.
00:05:45.360 --> 00:05:47.199
One is kind of changing culture.
00:05:47.199 --> 00:05:50.000
Culture is the hardest thing to change in an organization.
00:05:50.000 --> 00:06:14.879
And then I think there's that balance, particularly at the senior level, between doing leadership in a way that engages others and that you have a participatory and actually even somewhat antagonistic set of advisors, people who will give you disparate opinions about things, but at the end of the day, it's not a democracy, and somebody has to make the final call, and that is the CEO, and that's what you're getting paid for.
00:06:14.879 --> 00:06:23.680
And so, from my perspective, I always believe that there should, I should never get a decision across my desk that's a black and white decision.
00:06:23.680 --> 00:06:26.160
That should have been already made by somebody else.
00:06:26.160 --> 00:06:31.920
Every decision I ever made was shades of gray and required a judgment call of some kind.
00:06:31.920 --> 00:06:33.040
And that's the way it should be.
00:06:33.040 --> 00:06:37.519
When I inherited my organization, my predecessor was the exact opposite.
00:06:37.519 --> 00:06:39.279
He signed everything.
00:06:39.279 --> 00:06:44.800
And I was getting five dollar chits to sign off on so somebody could get lunch.
00:06:44.800 --> 00:06:46.319
We're not doing that here anymore.
00:06:46.319 --> 00:06:50.399
That's not quite what they hired me to do, is to sign off on a $5 lunch chit.
00:06:50.399 --> 00:06:56.000
And that that was a that's what that was a cultural change that required a huge amount of communication internally.
00:06:56.000 --> 00:07:00.800
We had quarterly all staff meetings where anything was fodder for conversation.
00:07:00.800 --> 00:07:05.759
And when people said we don't like that you're doing that, I said, I am open, give me alternatives.
00:07:05.759 --> 00:07:11.680
If you have them, I'm but you can't just say I don't want to do that without giving me a constructive alternative that is realistic.
00:07:11.680 --> 00:07:15.120
So there's lots that we could talk about in terms of leadership.
00:07:15.120 --> 00:07:27.600
But I think that that issue of engaging others to be part of the decision, but also being ready as a CEO, particularly, to be able to say, thank you for your input, but I'm not we're gonna go a different direction.
00:07:27.600 --> 00:07:28.879
You have to be able to do that.
00:07:28.879 --> 00:07:29.920
That's part of it.
00:07:30.240 --> 00:07:30.800
Absolutely.
00:07:30.800 --> 00:07:35.839
That is the hard part of the job of why CEOs get paid the most in most organizations.
00:07:35.839 --> 00:07:37.199
You have to make that hard call.
00:07:37.519 --> 00:07:43.839
As a consultant, I've been in organizations and it's been just incredible to me that people, the leadership that doesn't do that.
00:07:43.839 --> 00:07:52.959
It's like I used to think of organizations, particularly in the community behavioral health space, as being mom and pop shops at one end and the large sophisticated organizations.
00:07:52.959 --> 00:07:56.560
Now I think of it as the large organizations, sophisticated.
00:07:56.560 --> 00:07:59.040
And how in the heck are you still in business, folks?
00:07:59.040 --> 00:08:00.319
Because they're doing things.
00:08:00.319 --> 00:08:06.959
It's like, and largely it's the leader who doesn't want to take the leadership role and do the things that all the leaders can do.
00:08:06.959 --> 00:08:08.720
That's kind of the job of leadership.
00:08:08.720 --> 00:08:10.399
Do your job, make hard calls.
00:08:10.399 --> 00:08:12.399
So yeah, we could have a long conversation about that.
00:08:12.560 --> 00:08:12.879
Absolutely.
00:08:12.879 --> 00:08:25.199
And to go just a little farther, I think a lot of the a lot of the time where the mistake is in leadership is it's your job to make the hard call, to make the decision, and to know when you don't need to make the decision.
00:08:25.199 --> 00:08:29.040
To make the decision to trust your team to make decisions for you.
00:08:29.040 --> 00:08:31.600
I think that's probably the scariest place for most leaders.
00:08:31.600 --> 00:08:36.639
That's like the final boss to conquer, is now passing that responsibility on, but knowing you're still responsible.
00:08:36.639 --> 00:08:40.399
If your team makes a mistake, it's still on you, even though it wasn't your decision.
00:08:40.639 --> 00:08:56.320
Well, and it's interesting you mentioned that because I think that ties together what we were talking about previously, that the whole goal of augmented intelligence is to help clinicians do what only they can do and to get the other stuff off their desk, knowing that they are still responsible for it.
00:08:56.320 --> 00:09:06.159
And it's exactly the same process and leadership that if you were working for me, I have to say I have to trust that Tim can do his job and you contact me if you want some help.
00:09:06.159 --> 00:09:08.879
But but that's why I hired you to do your job.
00:09:08.879 --> 00:09:12.399
I should not know how to do your job better than you know how to do your job.
00:09:12.399 --> 00:09:14.720
I can help you, but that it shouldn't be that way.
00:09:14.720 --> 00:09:22.960
But that's I think that is one of the concerns that a lot of clinicians have about the idea that a good AI does not write the note for you.
00:09:22.960 --> 00:09:32.960
A good AI provides suggestions and ideas, insights that you may not have noticed, and you decide what goes in there, what does not go in there.
00:09:32.960 --> 00:09:43.360
But at the end of the day, it is the clinician, when they sign that clinical note, they are attesting to the fact that note is accurate and it represents what happened in that session.
00:09:43.360 --> 00:09:52.960
And that that I think a lot of clinicians, until they see how AIs should work, they tend to think, okay, this thing's sort of write my note for me and I'm not comfortable with that.
00:09:52.960 --> 00:09:54.399
No, it doesn't write your note for you.
00:09:54.399 --> 00:09:58.080
It gives you help so you can write the note faster, easier, better.
00:09:58.080 --> 00:09:59.759
But your point is exactly right.
00:09:59.759 --> 00:10:08.159
It's like the whole goal of this is to help people work at the top end of their license or at their jobs and get the detritus off of their desks.
00:10:08.399 --> 00:10:08.879
Absolutely.
00:10:08.879 --> 00:10:09.440
Exactly.
00:10:09.440 --> 00:10:15.919
So shifting topics to the last time here, since we're a marketing agency that works in this space, we always ask a couple of marketing type questions.
00:10:15.919 --> 00:10:19.679
And because you've had so many roles, I'm sure you have some good perspectives on it.
00:10:19.679 --> 00:10:25.200
But what have you found works best when it comes to acquiring talent, especially outside of word of mouth?
00:10:25.200 --> 00:10:28.559
What works best to attract good talent to an organization?
00:10:28.960 --> 00:10:37.039
Generically, I think if you create an environment that is attractive and that you clearly are helping people do their jobs better.
00:10:37.039 --> 00:10:41.919
I always thought my job was to help people do their jobs and to get barriers out of their way.
00:10:41.919 --> 00:10:46.000
It wasn't to fix things, but it was just to remove barriers they could not remove.
00:10:46.000 --> 00:10:48.639
But I think the culture has a lot to do with it.
00:10:48.639 --> 00:10:54.480
I mean, we meant word of mouth is still one of the primary ways in which people engage in organizations.
00:10:54.480 --> 00:11:07.279
On a more specific level, what we're seeing now is that people are using, in the case of Ilios, for example, where that I consult to, organizations are now using the fact that they're using Elios as a differentiator when they recruit.
00:11:07.279 --> 00:11:12.559
So they can go out and say, you know how hard it is to write your notes at your current organization?
00:11:12.559 --> 00:11:15.120
We've got a tool that can help you expedite that.
00:11:15.120 --> 00:11:21.279
And because there's a networking among clinicians, somebody will call a friend who works there and say, is that true?
00:11:21.279 --> 00:11:22.879
Is that hype or is that really true?
00:11:22.879 --> 00:11:24.080
And no, that's true.
00:11:24.080 --> 00:11:34.559
And it really does help literally to the point where we have heard on many occasions clinicians who have told their bosses, if you get rid of this AI tool, I'm leaving.
00:11:34.559 --> 00:11:37.679
Because it's helped them that much in their in their work.
00:11:37.679 --> 00:11:41.360
So people are using it as a differentiator in recruitment right now.
00:11:41.360 --> 00:11:42.080
I love that.
00:11:42.240 --> 00:11:42.639
I love it.
00:11:42.639 --> 00:11:44.639
That's a strong case for the service.
00:11:44.639 --> 00:11:54.799
Last question then is what is one marketing tactic or strategy that you or your team is either watching or just considering in the marketplace right now?
00:11:55.120 --> 00:11:59.200
Well, I don't know that we're there's any one that we're looking at to implement.
00:11:59.200 --> 00:12:02.000
There's one particular one, though, that's really got me worried.
00:12:02.000 --> 00:12:09.519
And that is, like every technology, you mentioned this, once it gets going, there's buzz and then there's knockoffs, and there's all kinds of things.
00:12:09.519 --> 00:12:36.639
And one of the things I'm one of the things I'm really concerned about is organizations, vendors that are bringing a generic AI like ChatGPT or something off the shelf that you and I would use for our kind of day-to-day work, putting a psychological wrapper around that and selling it as a psychology-related AI, when in fact it's nothing but ChatGPT or Claude or Perplexy or something on the back end, and they just put some kind of wrapper around it.
00:12:36.639 --> 00:12:54.080
That to me is worrisome because we know that, for example, ChatGTV chat was their large language model was trained on something between 16 and 30 percent of vetted content, books, magazines, and things like that, which many of them are being sued for copyright infringement, but that's another story.
00:12:54.080 --> 00:12:56.240
But the rest of it was just scraped from the internet.
00:12:56.240 --> 00:13:02.399
And that's why you and I can log into one of those off-the-shelf tools and ask it a general question.
00:13:02.399 --> 00:13:08.720
And it's very good because it's pulled information, but it's not psychotherapy, it's not what it's not what it's created for.
00:13:08.720 --> 00:13:12.480
It's a little bit like asking Siri to interpret an x-ray.
00:13:12.480 --> 00:13:15.360
Siri does a lot of fun things and it can be very helpful.
00:13:15.360 --> 00:13:16.399
That ain't one of them.
00:13:16.399 --> 00:13:23.519
And it's the same thing here where, you know, people they conflate just general conversation and interactions.
00:13:23.519 --> 00:13:25.279
That's the same as psychotherapy.
00:13:25.279 --> 00:13:30.639
And, you know, in AIs, their base models are trained to be sycophantic.
00:13:30.639 --> 00:13:33.919
They're trained to keep you engaged and to tell you what you want to hear.
00:13:33.919 --> 00:13:42.480
Well, that's caused a lot of problems, particularly for kids and adolescents who are using these off-the-shelf tools as a quasi-therapist.
00:13:42.480 --> 00:13:59.679
And there are many cases of kids who've committed suicide in concert while they're in conversation with these AIs, and they do nothing to refer them to 988 or to get them involved in some kind of crisis prevention or to dissuade them from thinking about committing suicide.
00:13:59.679 --> 00:14:07.679
So those are probably the biggest things that I see the idea of using AI in general in a way that it wasn't meant to be used for.
00:14:08.000 --> 00:14:08.559
Absolutely.
00:14:08.559 --> 00:14:09.600
I appreciate that.
00:14:09.600 --> 00:14:10.799
I appreciate your insights.
00:14:10.799 --> 00:14:13.360
I appreciate you taking time out of your day for us.
00:14:13.360 --> 00:14:16.639
Thank you, and thank you for the work that you're doing in your community.
00:14:16.879 --> 00:14:17.360
You bet.
00:14:17.360 --> 00:14:18.080
Thanks, Tim.
00:14:18.080 --> 00:14:18.799
Glad to be here.
