112 | Do you need an AI policy?

WEBVTT

00:00:00.017 --> 00:00:03.937
Hi there! Welcome to In the Clinic with Camille. My name is Camille Freeman.

00:00:04.097 --> 00:00:07.077
I'm an herbalist and nutritionist, and in this podcast I share tips,

00:00:07.177 --> 00:00:10.597
tidbits, and strategies, thoughts for other practitioners.

00:00:11.317 --> 00:00:14.977
Now, today I want to share with you something that I've been working on behind

00:00:14.977 --> 00:00:18.857
the scenes, which is a bloom-and-grow AI policy.

00:00:19.317 --> 00:00:22.577
You might have heard those last two words, the AI policy, and thought,

00:00:22.717 --> 00:00:27.497
oh my gosh, delete, skip, I can't even listen to another single word about AI,

00:00:27.837 --> 00:00:31.037
especially if you're somebody who doesn't choose to use AI yourself.

00:00:31.637 --> 00:00:35.077
It feels like you might just want to skip on through to the next one.

00:00:35.077 --> 00:00:39.957
I actually really encourage you to keep listening, though, because even if you

00:00:39.957 --> 00:00:44.117
personally are not working with AI right now, your clients are.

00:00:44.317 --> 00:00:48.497
If you are teaching, your students are working with AI, and therefore,

00:00:48.537 --> 00:00:53.757
you probably want to have some sort of written policy about this.

00:00:53.757 --> 00:00:56.677
So I wanted to share with you what I came up with.

00:00:56.817 --> 00:01:00.757
I've mentioned this, I think, on a few previous episodes, that this has been on my mind.

00:01:01.197 --> 00:01:05.537
Somebody who's really helped me to think through some of these is Amelia Ruby's

00:01:05.537 --> 00:01:10.677
Off the Grid podcast. She's had some really amazing guests on there and lots of food for thought.

00:01:10.897 --> 00:01:15.297
And a lot of what I came up with is based on things that I learned about or

00:01:15.297 --> 00:01:18.477
heard about or the work that they're doing over there on that podcast.

00:01:18.477 --> 00:01:22.337
So I'll put the link in the show notes to Amelia's, to the Off the Grid podcast

00:01:22.337 --> 00:01:25.197
and specifically some of the episodes that have AI policy links.

00:01:25.737 --> 00:01:29.757
What I've decided to do for right now is I've been working on a policy to put

00:01:29.757 --> 00:01:32.577
within the courses and programs that I'm teaching.

00:01:32.717 --> 00:01:38.737
And then I'm also working on my informed consent for clients that come to work with me one-on-one.

00:01:39.437 --> 00:01:42.697
I'm separately, and I'll do a different episode about this once it gets done,

00:01:42.877 --> 00:01:46.517
probably later in the year, in the fall or beyond. I'm separately working on

00:01:46.517 --> 00:01:51.257
a personal AI policy to publish, and I'll share with you a little bit the difference

00:01:51.257 --> 00:01:52.857
between those two things in a moment.

00:01:53.177 --> 00:01:57.857
But for right now, the policy that I've been thinking about is just how do I want to,

00:01:58.661 --> 00:02:03.641
how do I want to think about AI being used within the courses and programs that

00:02:03.641 --> 00:02:06.241
I'm running and with, let's just stop right there.

00:02:06.701 --> 00:02:09.881
So I thought on this for a while and I thought, you know.

00:02:10.814 --> 00:02:17.514
I feel that some people I know in my community are working with AI quite a bit.

00:02:17.934 --> 00:02:21.394
Other people have chosen to absolutely not have anything to do with it.

00:02:21.454 --> 00:02:24.034
And I understand both of those perspectives.

00:02:24.314 --> 00:02:28.434
I think it's really interesting and exciting technology. I think there's a lot

00:02:28.434 --> 00:02:31.674
of ways that it can streamline what we do and make us better clinicians.

00:02:31.674 --> 00:02:36.974
And I also understand the interest in not using it at all from an environmental

00:02:36.974 --> 00:02:42.174
perspective, from a creative thought perspective, from a saying as human as we can perspective.

00:02:42.174 --> 00:02:47.514
I think there's strong reasons to be on either end of the spectrum or anywhere in between.

00:02:47.514 --> 00:02:52.254
So what I really, really didn't want to do in my policy is make anyone feel

00:02:52.254 --> 00:02:56.894
judged for deciding what they've decided, or I didn't want to cut people off

00:02:56.894 --> 00:02:58.734
from feeling like they could change their minds later.

00:02:58.914 --> 00:03:01.554
Like, if you feel that you have to take a stand one way or the other,

00:03:01.654 --> 00:03:04.834
then it feels harder to change your mind as new information comes,

00:03:04.894 --> 00:03:09.454
as you try something out and it doesn't work, or as new tools and technologies

00:03:09.454 --> 00:03:12.334
become available, I think it makes sense to reevaluate.

00:03:12.854 --> 00:03:16.314
So anyway, I was thinking about all these things as I was coming up with my policy.

00:03:16.834 --> 00:03:23.614
I also was thinking, like, what do I value about the community that might be affected by AI tools?

00:03:23.894 --> 00:03:28.814
And how do I want to put some boundaries or make some requests around that?

00:03:29.494 --> 00:03:33.654
So let me just read you what I've come up with. This is a work in progress.

00:03:33.834 --> 00:03:38.234
I can't tell you that this will be exactly like it is for all time.

00:03:38.234 --> 00:03:44.774
But for right now, this is the community policy that I came up with for people in my programs.

00:03:45.294 --> 00:03:48.734
So it says, at Bloom and Grow, we are agnostic about AI tools.

00:03:48.994 --> 00:03:52.174
Some members use them liberally. Others have chosen to abstain completely.

00:03:52.454 --> 00:03:55.214
Because we're a community where sensitive information is discussed,

00:03:55.414 --> 00:03:58.094
including client cases and private information about business practices,

00:03:58.354 --> 00:04:03.034
we've developed the following policies related to generative AI tool use.

00:04:03.757 --> 00:04:07.377
Member AI assistants and note takers are not permitted to join Monday Mentoring

00:04:07.377 --> 00:04:09.417
or any Bloom and Grow live classes or sessions.

00:04:09.657 --> 00:04:12.877
We record all sessions via Zoom with closed captioning enabled.

00:04:13.077 --> 00:04:16.077
Recordings are available to all members within 24 hours. Okay,

00:04:16.137 --> 00:04:20.137
so let me pause right there and share a little bit about this piece of things.

00:04:20.457 --> 00:04:26.657
This is one of the places where it gets a little bit hazy for me because now

00:04:26.657 --> 00:04:32.557
there are note takers, there are AI note takers that do not notify you that they're joining.

00:04:32.557 --> 00:04:38.357
Back in the olden days, a year ago, two years ago, if an AI note-taker was joining

00:04:38.357 --> 00:04:41.457
the session, it would have to be admitted as a Zoom participant.

00:04:42.097 --> 00:04:45.217
You know, this is how Otter and some of these work, everybody's notified.

00:04:45.217 --> 00:04:48.877
You can see so-and-so's note-taking assistant is a member, you know,

00:04:48.937 --> 00:04:50.277
they're present in the call.

00:04:50.277 --> 00:04:54.757
But now things have gotten more advanced and there's plenty of note these ai

00:04:54.757 --> 00:04:59.157
note takers that you can just turn on and nobody else knows they're going so

00:04:59.157 --> 00:05:03.297
that was part of the part of the puzzle here for me that i felt like we needed

00:05:03.297 --> 00:05:07.617
to explicitly say this the other place where i felt a little bit.

00:05:08.339 --> 00:05:10.899
Not awkward, but I felt like I wanted to think about this intentionally,

00:05:11.179 --> 00:05:15.579
is I said member assistants or note takers are not permitted to join because

00:05:15.579 --> 00:05:18.299
Zoom has an AI assistant.

00:05:18.299 --> 00:05:25.299
I'm not using that right now, but we do use AI closed captioning and transcripting in Bloom and Grow.

00:05:25.479 --> 00:05:30.639
So I wanted to just say, I want to be in control of this for right now so that

00:05:30.639 --> 00:05:35.399
I can ensure that the data is being treated in a way that is, you know,

00:05:36.459 --> 00:05:38.599
legally compliant in case any

00:05:38.599 --> 00:05:43.079
client information is shared and that is not available to other people.

00:05:43.199 --> 00:05:47.139
So I think that's part of it for me is like, oh, if somebody's recording our

00:05:47.139 --> 00:05:51.179
calls without other people knowing and it's for their private use only,

00:05:51.359 --> 00:05:52.979
that really didn't feel right to me.

00:05:53.099 --> 00:05:58.219
I feel that when I am turning on AI transcripting or closed captioning,

00:05:58.339 --> 00:06:00.379
that that is available to everybody in the call.

00:06:01.119 --> 00:06:05.219
Specifically, it's helpful for folks who are, you know, for various reasons,

00:06:05.259 --> 00:06:09.019
people want and use those tools, but it's not just for one person,

00:06:09.159 --> 00:06:11.919
it's for everybody. And I'm using.

00:06:12.903 --> 00:06:17.483
Specific tools for specific reasons to do that. All right, so back into the policy.

00:06:17.663 --> 00:06:22.223
The next part of the policy says, generative or agentic AI tools may not be

00:06:22.223 --> 00:06:25.983
used to capture, scrape, or generate assets from any Monday Mentoring or Bloom

00:06:25.983 --> 00:06:29.403
and Grow content, including community discussion boards, slides and handouts,

00:06:29.563 --> 00:06:31.003
recordings, transcripts, and so forth.

00:06:31.403 --> 00:06:34.963
Please do not add Monday Mentoring or Bloom and Grow materials to large language

00:06:34.963 --> 00:06:40.323
models such as Claude, ChatGPT, or similar, even if you have data training turned off in your settings.

00:06:40.943 --> 00:06:44.443
Uploading our materials without consent of everyone involved violates our community

00:06:44.443 --> 00:06:46.383
trust and intellectual property boundaries.

00:06:46.863 --> 00:06:51.543
So again, this one is challenging because there's no way for me to control this.

00:06:51.683 --> 00:06:56.123
I do not know if somebody takes my recording or my slides or my reference list

00:06:56.123 --> 00:07:02.343
or what have you and puts it into Claude or whatever else, there's truly no way that I can stop that.

00:07:02.963 --> 00:07:07.203
And same as, you know, if somebody wants to screen capture my entire lecture

00:07:07.203 --> 00:07:11.623
video and save it for themselves to use later and cancel my subscription or whatever. Fine.

00:07:11.803 --> 00:07:15.223
That's something I don't have control over. But I did want to explicitly say

00:07:15.223 --> 00:07:17.123
that, hey, it's against my policy.

00:07:17.483 --> 00:07:21.363
Ethically, I don't think that's a reasonable thing to do, to take somebody else's

00:07:21.363 --> 00:07:27.423
work and upload it into one of these tools, because then it's out of our hands.

00:07:27.863 --> 00:07:31.483
Even if data training is turned off, they are still retaining that information.

00:07:31.883 --> 00:07:35.783
It's very hazy, the ethics of what's happening, the environmental impact, all of that.

00:07:36.203 --> 00:07:39.243
So this is what I decided. You may feel totally differently.

00:07:40.083 --> 00:07:42.943
And, you know, of course, any kind of notes people take on their own,

00:07:42.963 --> 00:07:44.463
they can do whatever they want with. But,

00:07:45.114 --> 00:07:49.774
For the things that I'm producing, I wanted to make it explicit that I do not

00:07:49.774 --> 00:07:52.694
consent to those things being added to large language models.

00:07:53.194 --> 00:07:58.894
Now, an interesting note, which I didn't address here, is that there are increasingly very,

00:07:59.214 --> 00:08:02.214
I don't know if they call them small, small language models,

00:08:02.434 --> 00:08:06.774
but there's some of these AI tools that can be hosted on your own computer that

00:08:06.774 --> 00:08:11.634
are for your own use only, that are not going beyond, you know, your little ecosystem.

00:08:11.634 --> 00:08:15.654
And I felt differently, I feel differently about those, but I feel like they're

00:08:15.654 --> 00:08:18.734
not, I don't know enough about them and I don't think enough people are using

00:08:18.734 --> 00:08:19.914
them at this point to really,

00:08:20.174 --> 00:08:23.654
I don't know, I didn't, I didn't feel that I could add that to my policy just

00:08:23.654 --> 00:08:26.894
yet, but I think for the large language models, this kind of covers it.

00:08:26.894 --> 00:08:32.454
Then the last piece of the policy says, disclose AI use within Monday Mentoring.

00:08:32.614 --> 00:08:38.354
If you use AI to help you draft a case study or generate protocol that you bring

00:08:38.354 --> 00:08:40.674
for feedback, please disclose that when you share it.

00:08:40.794 --> 00:08:43.294
There's no judgment and it is still helpful for us to know.

00:08:43.774 --> 00:08:46.654
Additionally, please keep community forum and chat interactions human.

00:08:46.854 --> 00:08:52.254
Do not use AI to generate replies or posts for the forum or in live chat during sessions.

00:08:52.734 --> 00:08:56.394
So I felt that was a pretty reasonable request. The reason that my programs,

00:08:56.514 --> 00:08:59.994
I think, are one of the things I really love about the programs is that we're

00:08:59.994 --> 00:09:02.274
together in real time and we're talking to each other.

00:09:02.694 --> 00:09:07.554
So I wanted to just make sure that we're keeping that present as much as we can.

00:09:08.448 --> 00:09:13.648
Okay, so there you go. This is what I've been working on behind the scenes for

00:09:13.648 --> 00:09:15.868
my courses and programs.

00:09:16.288 --> 00:09:21.288
Like I said, I'm doing a separate policy for how I use AI personally that I'm

00:09:21.288 --> 00:09:25.088
going to write up and put on my website. I am right now I'm experimenting with

00:09:25.088 --> 00:09:29.548
some tools and I may or may not decide to keep them in various ways.

00:09:29.808 --> 00:09:32.568
But I want to make that clear. I don't want to be using these tools without

00:09:32.568 --> 00:09:35.528
people understanding where and how and so forth.

00:09:35.988 --> 00:09:39.808
So I feel that in the interest of transparency, I'm going to create that, put it on my website.

00:09:39.808 --> 00:09:42.668
And I have just to give you a sneak

00:09:42.668 --> 00:09:46.328
peek one of the things that I've been experimenting with

00:09:46.328 --> 00:09:49.188
is doing writing up client

00:09:49.188 --> 00:09:54.308
notes not their protocols I do that myself still entirely and plan to do that

00:09:54.308 --> 00:09:59.008
forever but just taking a summary of the transcript of our call and putting

00:09:59.008 --> 00:10:05.008
it into a you know loose soap note format which is basically a data entry I've

00:10:05.008 --> 00:10:07.928
been experimenting with that But mainly from a time perspective where,

00:10:08.088 --> 00:10:11.448
you know, I don't feel like I'm adding any, it's not adding any new information

00:10:11.448 --> 00:10:15.508
to the call, but it is saving me probably 15 to 30 minutes of doing that myself.

00:10:16.496 --> 00:10:20.596
So I'm using a HIPAA compliant tool where that information is not being used

00:10:20.596 --> 00:10:22.376
to train any models and so forth.

00:10:22.556 --> 00:10:25.816
So I'm trying it out just to see how it works. I've added it to my informed

00:10:25.816 --> 00:10:30.196
consent disclosure for clients. They have the option to opt out of it if they want to.

00:10:30.556 --> 00:10:35.376
But that's one example of how I am trying some of these tools and so forth.

00:10:35.376 --> 00:10:39.136
So anyway, there you go. That's where I am right now.

00:10:39.296 --> 00:10:40.556
I've got this policy. I'm putting

00:10:40.556 --> 00:10:43.556
it up in various courses, trying to let people know that it's there.

00:10:43.776 --> 00:10:49.116
And then I am working on creating a more personal policy to put up on my website

00:10:49.116 --> 00:10:51.356
and experimenting with some things moving forward.

00:10:51.976 --> 00:10:55.156
All right. I hope that's helpful. I just I wanted to bring this to your attention

00:10:55.156 --> 00:10:59.636
because, again, if you are sharing educational materials, if you have courses

00:10:59.636 --> 00:11:02.996
that you teach, if you are sharing materials with your clients,

00:11:03.276 --> 00:11:06.556
you may want to make this clear for your own clients.

00:11:06.556 --> 00:11:10.956
Even if your whole thing is like, I don't use AI, I'm not interested in the AI, it's not part of it.

00:11:11.476 --> 00:11:15.256
But if you don't want your clients uploading your materials into AI,

00:11:15.636 --> 00:11:17.636
I think it's important to say that.

00:11:17.736 --> 00:11:21.216
If you've decided not to use AI, I think it's important to say that. Let people know.

00:11:22.116 --> 00:11:25.376
Some people, that might be a big selling point for why they want to work with

00:11:25.376 --> 00:11:28.576
you. Other people might be like, I want to work with somebody else. That's great.

00:11:28.796 --> 00:11:31.616
That's fine. It helps you bring the people, the right people into you.

00:11:31.616 --> 00:11:37.676
But I guess I'm just putting in a little plug for making your thoughts and your

00:11:37.676 --> 00:11:40.416
approach visible, and especially for...

00:11:41.252 --> 00:11:45.332
Letting clients know how you want them to, or students, what have you,

00:11:45.772 --> 00:11:51.532
what kind of permissions you're giving for your materials and what's possible to do with them.

00:11:52.432 --> 00:11:55.212
All right, I hope that was helpful. We'll talk about this more later.

00:11:55.212 --> 00:11:57.432
I'm sorry for adding to the AI discourse.

00:11:57.752 --> 00:12:01.832
I know there's just like a lot out there right now, and it can be kind of overwhelming.

00:12:01.972 --> 00:12:05.532
And a lot of us are like, I just want to be with the plants and in the kitchen.

00:12:05.992 --> 00:12:09.952
I don't want to be thinking about this or talking about it. But I think,

00:12:10.072 --> 00:12:11.172
you know, we have to think about it.

00:12:11.252 --> 00:12:14.072
We have to talk about it even if we choose not to use these tools ourselves.

00:12:14.512 --> 00:12:18.912
So there you go. I hope that's helpful. I'll link to some of my other episodes

00:12:18.912 --> 00:12:23.812
that have to do with AI in the chat as well as that in the off-the-grid podcast

00:12:23.812 --> 00:12:27.052
recommendation so that you can dive into that as well.

00:12:27.452 --> 00:12:30.112
All right, thank you so much for listening. I hope you have an amazing day.

00:12:30.732 --> 00:12:33.712
If you would like to stay in touch and you are not already signed up for my

00:12:33.712 --> 00:12:38.512
practitioner notes email newsletter, you can go to intheclinic.com or camillefreeman.com

00:12:38.512 --> 00:12:41.112
or the show notes to sign up to get that.

00:12:41.272 --> 00:12:46.752
It's always written by me, Never Any AI, homegrown newsletter for practitioners.

00:12:46.752 --> 00:12:49.932
It comes out every Thursday, and I would love to send it to you the next time

00:12:49.932 --> 00:12:52.532
it goes out. All right, I'll talk to you soon. Bye.