Latest / AI'm Just Trying My Best / Ep.7 : The Claude Watermark Story

Transcript

Ep.7 : The Claude Watermark Story · AI'm Just Trying My Best

• 0:02
  SPEAKER_00: Hello and welcome to another episode of AI, I'm just trying my best.
• 0:07
  SPEAKER_00: As always, I'm accompanied by my friend and business partner, Chevelle Pedra, principal AI strategist and founder of Modern Workery.
• 0:16
  SPEAKER_00: Chevel, how are you today?
• 0:18
  SPEAKER_01: I am doing amazing.
• 0:20
  SPEAKER_01: How are you?
• 0:21
  SPEAKER_00: I'm doing wonderful.
• 0:23
  SPEAKER_00: Thanks for asking.
• 0:24
  SPEAKER_00: Chevelle, I'm really excited to get into today's episode.
• 0:28
  SPEAKER_00: There's been uh a lot going on in the in the world of AI and interested to get your thoughts on some of these important topics that you know a lot of average people like myself might not fully understand.
• 0:41
  SPEAKER_00: So fortunately, you know, we're accompanied by you to uh help explain it to us a little bit more clear.
• 0:49
  SPEAKER_02: You are not average, you're great, and you know everything.
• 0:54
  SPEAKER_02: So smart.
• 0:55
  SPEAKER_02: What what Adam's making a little inside quip about is that I said he was an average user, which apparently did not sit well with Adam.
• 1:02
  SPEAKER_02: So he's a super user.
• 1:05
  SPEAKER_00: So Chamelle said you're an average user.
• 1:10
  SPEAKER_00: Wait a minute.
• 1:11
  SPEAKER_00: I don't want to sound offensive.
• 1:14
  SPEAKER_00: So usually when you follow back a statement with I don't want to sound offensive, you know that you sounded offensive.
• 1:21
  SPEAKER_00: So I will not take that to heart.
• 1:25
  SPEAKER_00: I will not hold that against you.
• 1:27
  SPEAKER_00: I still think the world of you and your knowledge, and I'm and I'm fortunate enough to call you my friend and to get to talk to you today.
• 1:34
  SPEAKER_01: Thanks.
• 1:35
  SPEAKER_01: Me too.
• 1:36
  SPEAKER_01: But um well, just me too, Adam.
• 1:40
  SPEAKER_01: Me too.
• 1:41
  SPEAKER_01: You're great.
• 1:43
  SPEAKER_00: Wow.
• 1:45
  SPEAKER_00: Anyhow, uh Chabelle, let's get into it.
• 1:47
  SPEAKER_00: We're gonna talk today about the Claude Watermark story.
• 1:51
  SPEAKER_00: And um, for those of you that don't know what that is and how it ties into, you know, Chabelle, can you give us a little glimpse of what the Claude Watermark story is?
• 2:01
  SPEAKER_02: Yeah, so Claude announced recently over the last couple of weeks that they are going to be introducing a watermark and its output.
• 2:09
  SPEAKER_02: And so it's for all new models going forward, and then they're actually gonna go back and retrofit the older models with this watermark kind of capability as well.
• 2:18
  SPEAKER_02: So everybody's freaking out about it.
• 2:20
  SPEAKER_02: Um, they don't know what that really means for them, how they're gonna be able to use, you know, uh, is it gonna impact them at work, at school, um, as a business owner, all the things.
• 2:31
  SPEAKER_02: And so it's uh it's a it's a big deal right now.
• 2:34
  SPEAKER_02: And it's it's part of um some of the EU legislation that's going on around AI.
• 2:39
  SPEAKER_02: So there's there's a lot going on.
• 2:42
  SPEAKER_00: Yeah.
• 2:44
  SPEAKER_00: And you know, we work so hard on getting business leaders to talk about AI, how we can teach them to use it, what's the best direction?
• 2:52
  SPEAKER_00: And then you have something like this break the news to where it's like, okay, what do we do?
• 2:57
  SPEAKER_00: How do we go back and backtrack?
• 2:59
  SPEAKER_00: How do we make things safe?
• 3:01
  SPEAKER_00: Right.
• 3:04
  SPEAKER_00: So how would you how would you explain to this like how you just mentioned um with with with with people that are that are integrating, you know, what would you say the biggest takeaway, the biggest concern should be with these watermarks that cloud's gonna roll out?
• 3:19
  SPEAKER_02: Yeah, so the EU legislation, so this is just one part of it.
• 3:22
  SPEAKER_02: And so what happened is these AI kind of hyperscalers companies kind of you know uh agreed to oblige and and you know, make sure that they're helping protect against things like deep fakes and you know, um, some of the bad stuff, like the bad actors that are using AI for not great purposes.
• 3:39
  SPEAKER_02: So the intent of the legislation is really, really positive.
• 3:43
  SPEAKER_02: And I think it's really sorely needed, quite honestly.
• 3:45
  SPEAKER_02: Um, and I think a lot of other company, you know, countries and and you know what are gonna follow suit.
• 3:50
  SPEAKER_02: But what happens is then that starts to kind of panic people of like, hey, I use you know, co-pilot for work when I'm writing emails.
• 3:57
  SPEAKER_02: Does that impact me?
• 3:58
  SPEAKER_02: And then they get scared.
• 3:59
  SPEAKER_02: And so I think there's a lot of um most of the the legislation that is in and you know coming in place and will be in place over the next few months, like don't actually impact employees.
• 4:10
  SPEAKER_02: Um, but I think it does bring up a greater story and an issue around people don't feel comfortable, they don't know when and how to use AI in the workplace, which I think is a bigger issue.
• 4:24
  SPEAKER_02: And I think something that you know companies really need to start to kind of tackle.
• 4:28
  SPEAKER_02: Um, because you have a lot of folks out there that are like use AI, but I'm scared to use AI.
• 4:33
  SPEAKER_02: Uh and that creates like a you know, a kind of a lot of unease among amongst people.
• 4:39
  SPEAKER_00: Yeah.
• 4:40
  SPEAKER_00: And so for average users like me, right?
• 4:44
  SPEAKER_00: And this isn't no, no, no dig or anything, but for average users like me, right, that say, okay, does this mean that every time I use AI to maybe write an email, to maybe create an image, there's gonna be a specific tag on it saying that I use AI to produce what you're seeing?
• 5:04
  SPEAKER_02: Yeah, so actually what's gonna happen is Claude and probably a lot of the other, you know, tools are gonna, it's it's essentially like metadata that's embedded.
• 5:13
  SPEAKER_02: So people are thinking it's gonna be some like watermark that's across the text or something like that.
• 5:16
  SPEAKER_02: That's not it.
• 5:17
  SPEAKER_02: It's actually metadata that's included into the text and also in the images.
• 5:21
  SPEAKER_02: Um, so if you generate an AI image, like there's gonna be metadata in that image that will, you know, indicate that it was used, you know, AI was involved in that process.
• 5:31
  SPEAKER_02: Um the the data is not collecting who used it, you know, to what extent it was used, um, no type of data on the end user.
• 5:40
  SPEAKER_02: So from like a privacy perspective, that there's nothing there right now to worry about.
• 5:46
  SPEAKER_02: Um, but it will indicate and tools will be able to detect whether AI was a part of that process.
• 5:53
  SPEAKER_02: Um and so that's what people are kind of freaking out about.
• 5:56
  SPEAKER_02: Um and so from a EU perspective, from a legislation perspective, if you're doing business with the EU, it doesn't really impact knowledge workers, right?
• 6:05
  SPEAKER_02: Like so if you're just kind of using co-pilot for personal productivity within an organization, none of it really, you know, applies to you necessarily right now.
• 6:14
  SPEAKER_02: Um now if you're interfacing directly with customers using AI, that might be where there's some some kind of legislation that comes into play.
• 6:22
  SPEAKER_02: But it it just, you know, it's scaring people because they're like thinking that they're gonna like send out an email that they use co-pilot for and there's gonna be like a big red like letter on that email, like letting everybody know.
• 6:35
  SPEAKER_02: It's kind of like the M-dash, right?
• 6:36
  SPEAKER_02: Um, people are always trying to hide those M-dashes that just it's kind of the same concept.
• 6:43
  SPEAKER_00: So, an important question I know all my my college uh college students are gonna want to know is is this going to be something that they're going to have to work up against with helping maybe publish papers or do research, right?
• 6:58
  SPEAKER_00: Is this something that's going to help professors or help uh individuals identify what's been created by AI versus what hasn't been?
• 7:06
  SPEAKER_02: Yeah, I mean, there's AI detection tools right now that universities have been using since AI basically was was kind of commonplace.
• 7:14
  SPEAKER_02: Now they're not typically great.
• 7:16
  SPEAKER_02: Um, and so I don't know what this is gonna kind of end up looking like from an education perspective.
• 7:22
  SPEAKER_02: I think they're still figuring all that out.
• 7:25
  SPEAKER_02: But I think the important caveat here is that it doesn't like distinguish how much you used AI, right?
• 7:34
  SPEAKER_02: So you could use it to have it review your content and there's, you know, and just kind of provide some edits and things like that, which is a perfectly permissible.
• 7:45
  SPEAKER_02: I think everybody would agree that's a great use case for AI.
• 7:48
  SPEAKER_02: There still may be those watermarks within that text.
• 7:51
  SPEAKER_02: Um, and so, or you could just create the whole thing, you know.
• 7:55
  SPEAKER_02: So it's gonna be interesting to see how that, you know, plays out a bit.
• 7:59
  SPEAKER_02: Not that you would ever have uh claud write your college papers, um, but um some people might.
• 8:06
  SPEAKER_02: So we'll see how that that plays out.
• 8:08
  SPEAKER_02: Um but I think the biggest thing that folks need to just know is the main kind of criteria for whether you need to disclose that you are using AI for a purpose is human in the loop.
• 8:21
  SPEAKER_02: So if you were, you know, actively engaged in the creation of the content, you were actively engaged in the editing, like you reviewed it, you know, is a really collaborative effort with AI, that doesn't necessarily have to be disclosed.
• 8:35
  SPEAKER_02: Um, it's when you just kind of have the AI create content, you know, don't look at it, just have it go out.
• 8:43
  SPEAKER_02: Like that's where the transparency requirements are going to come in as well.
• 8:46
  SPEAKER_00: Gotcha.
• 8:48
  SPEAKER_00: And you know, for our listeners out there, you you hear EU AI Act, right?
• 8:53
  SPEAKER_00: And they say, well, that's in Europe.
• 8:55
  SPEAKER_00: That doesn't really affect me in the United States here.
• 9:00
  SPEAKER_00: Is that a true statement or a false statement?
• 9:02
  SPEAKER_00: And how does the two collide?
• 9:04
  SPEAKER_02: No, I mean, I think all companies should start really paying attention to this legislation because it's even if you do business with anybody in the EU, right?
• 9:13
  SPEAKER_02: And which a lot of people, you know, do business with people in the EU.
• 9:17
  SPEAKER_02: Or, you know, say you're a content creator or you have a website, somebody could go on your website.
• 9:23
  SPEAKER_02: Um, you can't control where they come from, you know, so so that would be a requirement for you as well.
• 9:29
  SPEAKER_02: So I think it's um similar to like other, you know, prior EU legislation.
• 9:34
  SPEAKER_02: I think you'll see that the larger enterprises start to kind of follow suit and adhere to this.
• 9:39
  SPEAKER_02: Um, but I think it's it's good practice for everybody to start to get really familiar with what this means for them um and and start to incorporate you know some of this into their AI policies and practices.
• 9:53
  SPEAKER_00: So yeah, and and I I I couldn't agree with you more because you know, when we look at we see a lot of articles about what other countries are doing to kind of regulate AI, right?
• 10:06
  SPEAKER_00: And we don't see as many legislation acts in the United States as we do in other countries.
• 10:13
  SPEAKER_00: Is there a reason why, or that you think that other countries are a little bit more progressive when it comes to um honing in on uh keeping you know AI within a standardization so that it doesn't hurt or affect people?
• 10:30
  SPEAKER_02: Yeah, I mean, I think the EU's always been a little bit ahead in terms of things like the GDPR and and things like that.
• 10:39
  SPEAKER_02: Um then the US.
• 10:41
  SPEAKER_02: Um, and I think we're gonna find that up to this point, the the technology is evolving so quickly, it's kind of impossible to keep up.
• 10:50
  SPEAKER_02: So these, you know, these hyperscalers are kind of governing themselves a bit, you know, because they're so far advanced, then where the legislation and you know the capabilities of the government to even understand what's happening and how to, you know, um keep their arms around it and whatnot is.
• 11:08
  SPEAKER_02: So um, you know, not to say that in the future, you know, we we might have more kind of aggressive stances against it or or for it, but you know, typically the EU's ahead of the game on these.
• 11:23
  SPEAKER_00: Where do you see the high risk systems that would be included in this be?
• 11:27
  SPEAKER_00: Like what would you say would be the start of you know, kind of following the same the same system of as as watermarks within not only Claude, but with other AI or LLMs?
• 11:38
  SPEAKER_00: Like, what do you see the high risk stakes here within within operating within watermarks and letting the public know or letting the people know, you know, what's been created by AI and what hasn't been?
• 11:49
  SPEAKER_02: Yeah, so the the legislation talks a lot about like high risk activities.
• 11:55
  SPEAKER_02: And so that is your hiring, firing, you know, credit decisions, like anything that you're like, well, yeah, people really probably need to be involved or at least informed um when you know there is AI involved in that process.
• 12:07
  SPEAKER_02: So anything that you're like, eh, you know, it's kind of weird.
• 12:11
  SPEAKER_02: Right?
• 12:11
  SPEAKER_02: I just become on your own.
• 12:13
  SPEAKER_02: Yeah, that's it's probably a high risk activity.
• 12:15
  SPEAKER_02: Um, and so I for again for most people, like probably listen to this podcast and you know, that that will run upon this, it's not really important to know the you know details of that.
• 12:28
  SPEAKER_02: What is more important though is for companies, and I think this is the bigger issue right now, especially, is that employees just don't know what they're allowed to use it for, what they should be using it for.
• 12:41
  SPEAKER_02: And a lot of companies are starting to have AI policies, which start to kind of govern some of this the use, but what are the norms, right?
• 12:50
  SPEAKER_02: Like you don't want to be that employee that your whole team doesn't use co-pilot.
• 12:55
  SPEAKER_02: Well, you use co-pilot to draft your, you know, you kind of have a bit of shame employees might have about using that.
• 13:02
  SPEAKER_02: I think that's the bigger issue right now, quite honestly.
• 13:05
  SPEAKER_02: I mean, Adam, have you ever seen that where you're like, I don't, I try to, you know, you hide the M-dashes or you try to hide, like put a typo in there just to make it seem like it's you versus AI using it.
• 13:19
  SPEAKER_00: Yeah, I see that a lot, right?
• 13:21
  SPEAKER_00: I see that a lot in um, you know, where we talk about when we speak with leadership, right?
• 13:27
  SPEAKER_00: And and we say, hey, you should be utilizing, you know, your your co-pilot for more than just checking emails, right?
• 13:33
  SPEAKER_00: Or more just um prompting it to rewrite an email to sound more professional, right?
• 13:38
  SPEAKER_00: We're trying to help utilize, uh help organizations utilize co-pilot within a higher functionality.
• 13:44
  SPEAKER_00: However, excuse me, it's still the same thing where you read an email and you're like, oh god, AI wrote this for sure, right?
• 13:51
  SPEAKER_00: It's almost an internal game you play with within your head.
• 13:54
  SPEAKER_00: It's either did AI write this or did it not, right?
• 13:57
  SPEAKER_00: Yeah, and it feels like you you I struggle with that.
• 14:00
  SPEAKER_00: I'm like, oh, this person used AI for sure.
• 14:02
  SPEAKER_00: This isn't them writing it, you know.
• 14:04
  SPEAKER_00: And even to the point where, you know, I think we discussed it last week how you know I'm uh taking my son to football practice, and there's a sign, you know, stuck out in the corner of the intersection, and he tells me, holy AI.
• 14:19
  SPEAKER_00: And I'm like, holy AI, what's that mean?
• 14:21
  SPEAKER_00: He's like, Oh, that means they did they they used AI to complete that whole thing, right?
• 14:25
  SPEAKER_00: And it's almost looked at as a as frowned upon, like you've lost your creative, uh your creativeness to just being lazy, is what it feels like, right?
• 14:35
  SPEAKER_00: And and that's kind of instead of those games where you played.
• 14:38
  SPEAKER_00: I don't know if you remember when you traveled when you were a kid, you play those car games where you look for license plates or um you know slug bug, right?
• 14:47
  SPEAKER_00: Now it's almost like you're identifying advertisements on billboards, whether they're created by AI or not.
• 14:53
  SPEAKER_00: You know, it's either wholly AI or no, that's actually real.
• 14:56
  SPEAKER_02: Yeah.
• 14:57
  SPEAKER_02: And if you're really nerdy, you can kind of start to pick up on what model they use to write that email.
• 15:03
  SPEAKER_02: Like, oh, that's fable.
• 15:04
  SPEAKER_02: That's a fable one for sure.
• 15:06
  SPEAKER_00: Um, yeah, a lot of average users like myself don't pick up on that stuff.
• 15:11
  SPEAKER_00: So we haven't got to that game yet.
• 15:14
  SPEAKER_01: Oh, I'll never live that down.
• 15:16
  SPEAKER_02: Um but I mean, I think so, but it I thought that, right?
• 15:22
  SPEAKER_02: Like I'm an AI like adoption specialist expert.
• 15:26
  SPEAKER_02: And I still sometimes am like, oh, this sounds way too professional.
• 15:29
  SPEAKER_02: I gotta make this sound more dumb.
• 15:30
  SPEAKER_02: Some people think it's me.
• 15:31
  SPEAKER_02: Um, and so that's like the complete opposite of what these tools are designed for.
• 15:36
  SPEAKER_02: And that's where I think the biggest thing that companies need to do right now is create norms and create a list of when AI is expected to be used, when AI should support, and when it should be human first.
• 15:51
  SPEAKER_02: So like three buckets, right?
• 15:53
  SPEAKER_02: And and varying levels of AI support.
• 15:56
  SPEAKER_02: And I think well, I have uh go ahead.
• 16:00
  SPEAKER_00: No, I was gonna say I couldn't agree with you more.
• 16:02
  SPEAKER_00: And you mentioned earlier, you know, when you said high-risk systems, uh, with letting people know that AI is being used.
• 16:09
  SPEAKER_00: Uh, you said one was hiring, right?
• 16:12
  SPEAKER_00: And what I'm seeing a lot now when I talk to a lot of my HR colleagues, is what does that look like from adoption in HR for recruiting agents and losing that human touch?
• 16:25
  SPEAKER_00: And I couldn't help but think of when you mentioned the hiring aspect of it, is right now you're seeing a lot of job seekers just be disgruntled, and or um when they apply to jobs, they're applying to hundreds of jobs and getting no feedback.
• 16:41
  SPEAKER_00: And there's a very negative stigma about AI within hiring processes because they don't feel like they're actually talking to anybody, they're feeling that their resume is going into an agent or uh a chat bot, it's not being reviewed by somebody, it's being reviewed by an agent that's listened, that's that's built to recognize terminology, right?
• 17:03
  SPEAKER_00: And I feel like companies are gonna lose to that, lose good people to that because they're taking the humanistic trait out of the hiring, which is hiring for you know character, hiring for skills, abilities, right?
• 17:16
  SPEAKER_00: But um, I feel like that's gonna help the job market, help the job seekers find, uh apply to positions that they know are not rained by AI if there is a watermark for it.
• 17:27
  SPEAKER_02: Yeah.
• 17:27
  SPEAKER_02: And I mean, if you think about that's a good example, because if you think about the the day of a recruiter, there's probably so many activities that have nothing to do with a human or any type of human judgment, you know, um, writing status updates, you know, summarizing meeting notes, you know, writing this, whatever it may be.
• 17:46
  SPEAKER_02: Like there's so many of those moments and and tasks that a recruiter has to do on a given basis that takes them away from actually like reviewing resumes, from having more, you know, kind of, you know, uh deeper conversations with folks.
• 17:59
  SPEAKER_02: It's for like use AI for for that stuff, the work around the work, not the work itself.
• 18:06
  SPEAKER_02: And I think to start again, I think it it helps companies by like just sitting down and again having a list, having a on the whiteboard, whatever it is, like those three categories.
• 18:18
  SPEAKER_02: You know, AI is expected, AI supports human first, and just starting to map out like where that work is.
• 18:26
  SPEAKER_02: Um, and I think the biggest thing is is like to make sure for AI expected, um, that everybody starts to use AI for that purpose.
• 18:37
  SPEAKER_02: And not because we're trying to force people to use AI, but where I think you'll start to really get a shift is again, you can say, hey, everybody in the company should use AI.
• 18:47
  SPEAKER_02: But again, if you're like not the only one, if or if you're the only one using AI, you're still gonna feel like, should I be using it?
• 18:54
  SPEAKER_02: Should I not?
• 18:55
  SPEAKER_02: So maybe that 10 minutes you save on having co-pilot draft you an email, if you spend 15 trying to figure out how to make it not sound like co-pilot wrote that email, like what are we doing?
• 19:08
  SPEAKER_02: And I think there's a whole lot of that out there.
• 19:12
  SPEAKER_02: Um, and they think the importance of having expectations of when to use AI is kind of like, you know, the NFL, right?
• 19:18
  SPEAKER_02: You have those guardian caps.
• 19:20
  SPEAKER_02: They I think they look better this year, but like last year they looked kind of silly.
• 19:24
  SPEAKER_02: They had like a bunch of padding and they helped prevent concussions, right?
• 19:29
  SPEAKER_02: I don't think like anybody says concussions are great.
• 19:32
  SPEAKER_02: I think NFL players should get more concussions, but nobody wants to wear the helmets because they look stupid, like they feel like they're isolated.
• 19:43
  SPEAKER_02: If everybody wore the damn guardian caps and were required to use the guardian caps, it wouldn't be a big deal because everybody was it just becomes the norm.
• 19:53
  SPEAKER_00: Yeah, yeah.
• 19:53
  SPEAKER_02: Like everybody, you know, everybody's it's it's better.
• 19:57
  SPEAKER_02: So it's like status updates.
• 19:59
  SPEAKER_02: Like everybody just send their status updates with AI.
• 20:03
  SPEAKER_02: AI is probably gonna do a better job than you writing that status report.
• 20:06
  SPEAKER_02: Now you of course review it always and make it your own, but like if everybody was doing it, that wouldn't be a big thing.
• 20:13
  SPEAKER_02: Um, and a lot of companies don't sit down and and really dictate like those things that employees should be using AI for.
• 20:21
  SPEAKER_02: They're focusing a whole lot on what they shouldn't use it for, but not on what they should.
• 20:25
  SPEAKER_00: Yeah, I feel like I just like they need to play a game AI or not AI, right?
• 20:31
  SPEAKER_00: So you get your whiteboard, you put one section AI, draw a T, not AI, write job descriptions, AI, conduct an interview, not AI, yeah, right?
• 20:42
  SPEAKER_00: And just kind of go through the simplistic form of keep the humanistic traits human and use AI to navigate those those hassle things, the administrative things, the data things that you need to do, right?
• 20:55
  SPEAKER_00: AI, not AI.
• 20:56
  SPEAKER_00: It's pretty simple, yeah.
• 20:58
  SPEAKER_02: It is, and a lot of companies focus so much on like these big fancy AI policies and governance structures and things like that.
• 21:06
  SPEAKER_02: But and then they have all these fancy dashboards about adoption, but it's like your teams, I promise you, do not know what is acceptable.
• 21:13
  SPEAKER_02: It's kind of like the whole norms, you know, um, of a team.
• 21:17
  SPEAKER_02: Like, what is acceptable to use it?
• 21:19
  SPEAKER_02: So they're either not using it, which is not what you want if you're paying for the tool, or they're using it and figuring out ways to make it seem like they're not using it, which is also not productive.
• 21:30
  SPEAKER_02: So simple exercise, but it'll save a ton of time.
• 21:34
  SPEAKER_02: And then it will help you identify those high-risk areas that maybe AI should be more of like a support function, more of like a thought partner.
• 21:42
  SPEAKER_02: Um, but just start clipping away at like, hey, what are those stupid things we do every day that AI can just take off our plate, you know?
• 21:50
  SPEAKER_00: Yeah.
• 21:51
  SPEAKER_00: And I think as leaders need to think about it, simplicity is the best, right?
• 21:54
  SPEAKER_00: Sometimes a less is more.
• 21:56
  SPEAKER_00: And when we talk about um implementing a strategy.
• 22:00
  SPEAKER_00: Strategy, right?
• 22:01
  SPEAKER_00: Simplify it.
• 22:02
  SPEAKER_00: Make a whiteboard saying yes, AI, no AI, what do we want to use AI for?
• 22:06
  SPEAKER_00: And then start mapping it out.
• 22:07
  SPEAKER_00: Because the last thing you want to do is invest in all of these tools, create all these dashboards that are confusing, that when you start to implement it within your workforce, you know, people's eyes go crossed and they are given up before they even start the race.
• 22:23
  SPEAKER_00: And that that can be, you know, detrimental to your work staff, right?
• 22:27
  SPEAKER_00: And then next thing you know, you're paying for the tools, you're not utilizing them, and then you start all over again.
• 22:32
  SPEAKER_02: Yeah.
• 22:32
  SPEAKER_00: Yeah.
• 22:33
  SPEAKER_02: And you know, it's funny.
• 22:35
  SPEAKER_02: I when I read about the the cloud watermark and all of this, um, it reminded me I used to have a pretty uh hearty debate with one of my peers who was actually in charge of rolling out co-pilot to a large organization.
• 22:48
  SPEAKER_02: Like they were, you know, if anybody's using AI, it should be like the people rolling out, you know, into the rest of the organization.
• 22:56
  SPEAKER_02: And there is an email, and they're like, oh, so-and-so clearly used AI, almost like it was a bad thing.
• 23:02
  SPEAKER_02: I was like, well, cool.
• 23:04
  SPEAKER_02: Like, isn't that the point?
• 23:05
  SPEAKER_02: And they're like, no, that's not the point.
• 23:07
  SPEAKER_02: Like, and and they were right, and I was right, you know, like there's there's certain emails, even like based on subject.
• 23:15
  SPEAKER_02: If it's a stupid status report, yeah, like have an executive write, use AI.
• 23:18
  SPEAKER_02: But if it's like, hey, we're going through a tough time as an organization, we might have you know, we might have face some some tough, like you don't want AI writing that.
• 23:27
  SPEAKER_02: Like, so that's where I think common sense also comes in.
• 23:31
  SPEAKER_02: Um, right, but even kind of breaking it down and like what do we like from a subject matter?
• 23:36
  SPEAKER_02: Like, when is it appropriate for AI to kind of take the wheel?
• 23:39
  SPEAKER_02: And when should we take the wheel?
• 23:41
  SPEAKER_02: Um, you know, those are really valuable conversations worth having up front.
• 23:45
  SPEAKER_02: It saves a lot of heartache, a lot of turmoil, a lot of, you know, just you know, unproductive time later down the line.
• 23:54
  SPEAKER_00: Yeah, make it a could you imagine getting a notice of termination with an AI generated email?
• 24:03
  SPEAKER_02: I guarantee it's happening all day, every day.
• 24:07
  SPEAKER_00: You want to talk about horrific exit interviews.
• 24:11
  SPEAKER_00: Uh, use AI to generate a termination email and then send it to those people being terminated.
• 24:17
  SPEAKER_02: I guarantee it is happening, it's happening like right now.
• 24:20
  SPEAKER_02: I guarantee it.
• 24:21
  SPEAKER_02: Um and I I I guarantee the in executives probably applauding themselves for using, they probably created an agent for it.
• 24:29
  SPEAKER_00: Um could you imagine what you what'd you use AI for?
• 24:33
  SPEAKER_00: Well, I created an agent to generate termination emails so that I can make them non-human but human realistic, semi-human to make it seem like you know, I gave a damn about you, but in in the end, I didn't.
• 24:46
  SPEAKER_02: I know those agent instructions.
• 24:48
  SPEAKER_02: Um that's you know, and it's funny, people just go to these use cases.
• 24:55
  SPEAKER_02: I guarantee.
• 24:56
  SPEAKER_02: So let's let's let's play this out in that executive's day.
• 25:00
  SPEAKER_02: There's probably so many things that they could have used AI for that would have given them the time to sit down and write a more thoughtful like termination letter.
• 25:09
  SPEAKER_02: Um so it's like people aren't like, you know, they're just going for this.
• 25:13
  SPEAKER_02: Sometimes it's just use a little bit of common sense and empathy and like judgment.
• 25:18
  SPEAKER_02: Um it'll it'll typically always work well when you do those things.
• 25:23
  SPEAKER_00: That's the solution right there.
• 25:25
  SPEAKER_00: Common sense, empathy, and good judgment.
• 25:28
  SPEAKER_02: You know, I I use that those things sometimes.
• 25:31
  SPEAKER_02: Um not all the time.
• 25:34
  SPEAKER_02: I try.
• 25:35
  SPEAKER_02: I'll ask Copilot to ask me how to be more empathetic and use better judgment.
• 25:40
  SPEAKER_00: Um, you know, whether we're trying to be uh better at at using better judgment, better at AI agents, as long as you're not utilizing them to lay off your workforce, I think you're in a good spot.
• 25:55
  SPEAKER_01: Yes, I agree.
• 25:57
  SPEAKER_02: Use always grow your business, figure out how to grow your business using AI.
• 26:03
  SPEAKER_00: Yeah, yep.
• 26:04
  SPEAKER_00: As always, Chabelle, it's been a pleasure.
• 26:06
  SPEAKER_00: Thanks for uh chatting with me today, and we'll talk to you next week.
• 26:09
  SPEAKER_02: Awesome.
• 26:10
  SPEAKER_02: Thanks, Adam.
• 26:10
  SPEAKER_02: Take care.
• 26:11
  SPEAKER_02: And you're not just average, you are uh very above average.
• 26:15
  SPEAKER_02: If anybody has not told you that today, thank you, Chevelle.
• 26:19
  SPEAKER_00: That means the world.
• 26:21
  unknown: All right.
• 26:22
  SPEAKER_00: Until next week, we'll see you then.
• 26:23
  SPEAKER_00: Bye.
• 26:24
  SPEAKER_00: Bye.