Navigating the Future of AI & Cybersecurity | TSG Ep. 892

Can you believe it? 25 billion. What are we talking about?

We're gonna touch on that in a lot more. You're joining Textron Gang. Welcome.

So glad you joined us today for the Textron Gang. We've got our wonderful member of, uh, panel of gang members, uh, some folks that you probably know. We have one newcomer, uh, John Schwartz, of course, who writes, uh, for Techstrong in many of our publications.

Uh, g who I've known from DevOps and lots of conferences and very involved in hackathons. Terry, who also writes a lot for us as well. And Jeff Rich.

Jeff, tell us a little bit about you. Give us a couple sentences on who you are. Well, Mitch, I'll, I'll, I'll pick just a couple.

Um, I, I lead the Identity Defined Security Alliance. We are a nonprofit, um, organization focused on raising a level of identity and identity security. Uh, we have a, a lot of, uh, members that I'm sure listen to this, and I think we're gonna talk about one of those members in just a couple minutes.

May just do that. So, and tib by the way, Terry does a lot more than just write things for her. She's, uh, definitely a thought leader in cybersecurity, so thank you, Terry.

Hey, before we kick things off, um, we want to touch on sort of the, the giant elephant or big gorilla, or whatever it is in the room. Uh, who is it? Palo Alto spent a little about Palo Alto Networks, just literally a few minutes ago, just said.

They, they have agreed to acquire CyberArk software for about $25 billion, which would be the biggest deal yet for Palo Alto Networks in an industry that is consolidating somewhat in cybersecurity. There are a lot of deals going on as AI propels demand among enterprises. So this is a cash in stock deal.

It was just announced, it's scheduled to be, to close in fiscal 2026. It just happened. We're gonna talk more about this on Friday show, but I, I think Jeff, um, you alluded to knowing some of the participants or particulars, maybe you could just weigh in a little bit briefly what, on, on what you think this means, and then we'll move on to our regularly scheduled program.

Sure. And, and thanks John. I, I, I'm not gonna talk about the financials.

Hmm. CyberArk is a board member at IDSA. And, uh, they've been very active, done a lot of innovation in the identity space in particular.

And I think what this, um, acquisition really shows more than anything is that as each day moves forward, we're seeing security and identity not only have a parallel path, but in many cases have this the same path. And, and I think this, um, identifies that when you look at, you know, identity and access management as an identity feature, and then say, what effect does that have on security? I think mergers like this are gonna start telling us that you're gonna know what effect it has on security, because they will eventually come one and the same.

Okay. Okay. Yeah, there, there was just the, uh, I think it was in March, alphabet, um, announced that three, $2 billion acquisition of Wizz, which is their biggest deal too.

So I think more is to come, and we will talk more about this on tomorrow's show, but I just wanted to acknowledge the, uh, the gorilla in the room. Now let's move to the A block and AI skills required. Um, there was a DevOps for Gen AI hackathon recently in Ottawa that I believe Garima that you attended, or that you, you, I know you wrote about it.

I think you attended it. And I wanted to, to, to kind of scan your mind on, on what the highlights were, um, the growing convergence of AI observability and prompt engineering and projects list between two domains and things like that. I'm wondering kind of what your takeaways from this, this hackathon were.

Thank you for this question, John. And this is very interesting development. I would start with some industry insights, uh, to build some credibility here.

So about 30% of our work activities as software practitioners, software developers, and, you know, whatever paradigm shift we are seeing in the software world could be automated by 2030, uh, accelerated by Gen AI deployments. And this is a report from McKenzie. So you know what essentially it means for practitioners and organizations.

So I'll start with like practitioners, software developers. Uh, there will be new skills required for this new era and new age of ai, right? We'll have new teammates like AI coding boards, coding companions.

So we'll have to be ready for that. So as a team, we have to embrace that change for, um, executives and for, uh, you know, senior management. What it means, this shift or change is that, uh, there is an async first culture, which is building.

So we'll have to collaborate with ai, uh, in a broader way. We will also have to embrace AI native workflows, right? And to a large extent, I mean, this is, again, my take on this is that for executives, it is more or less, uh, about sustainability, responsible injection of AI into software development practices, and how do we build ethical practices around it, right?

So, having said that, this looks all fancy, but what practically would happen in organizations is that we will need a lot of, uh, new persona types, right? So if you start from the bottom of, uh, the chain, and, uh, again, the, the most important and the most, uh, the, the core component of this would be programmers, right? So I would call them AI programmers.

So they are leading that shift and the innovative approach for software development, right? So we will need more and more of that skillset. We also would need prompt engineers.

I mean, I, back in time, I thought that it was not a thing, but you know, right now, I think what we, uh, are seeing is effective prompt writing is becoming a, a, a, a, a kind of skillset. And how, uh, we would embrace that shift. We also would need, like AI specialists, researchers, and to the core of this is also data engineers.

Like, how do we clean data? How do we access data? How do we make our data ready for AI models?

So data engineers will play a important role in this. And lastly, I would say we would need data scientists and machine learning engineers for developing the capacity and deploying these models. So obviously there is a substantial amount of shift happening as we speak.

So when, um, I was at a conference, uh, last year, and we were brainstorming this, um, and had happened to meet John Willis at that conference. If, uh, uh, people don't know about John Willis, I mean, uh, you must be hiding beneath the rock. But, uh, he is one of the visionaries in the DevOps site.

That's One of the godfathers of DevOps, or it's, exactly, Exactly. So we actually, we brainstormed, uh, a lot of ideas that how do we build capacity for all this, right? I mean, this is not that, you know, tomorrow morning we'll wake up and as executives, as senior leaders, as, uh, you know, practitioners, we will find these skill sets.

So we will have to embrace that change. And how do we develop this capacity is through, you know, hackathons. And this was a great idea, which John brought forward, and we actually thought about it that, okay, uh, let's, uh, put this, uh, hackathon in place.

And this was one of the first series which happened in Ottawa. So DevOps for Gene AI hackathon, uh, was kind of presented. And, uh, it was, I mean, I'll tell you, it was, it, it's not only about coding anymore, these hackathons, I mean, it's about, uh, building a vibrant community where industry professionals, academy students, uh, come together and we bridge the gap between theory and practice, and also fueling that innovation, which is required where we can scale prototypes into production ready solutions.

So, giving you some, some examples, what happened at the hacker Hackathon was like, people started to imagine a world where AI deployments monitor themselves in real time automatically and detect issues like hallucination or optimizing from quality through continuous feedback loops, and even, uh, to a certain extent, self-heal, uh, with minimum downtime. So there were three projects which stood out inside AI minions and inner AI middleware and gen AI helper. I mean, putting these three projects together, I mean, I would say that, uh, we would be seeing a lot of like innovative, uh, solutioning if, let's say open source community comes together to develop these projects, right?

So it, this, this is the kind of potential of these hackathons. Um, we also understood that, uh, despite the promising approach, which we took, uh, with this hackathon, I think we also understood the challenge, like deep DevOps integration, for example, enterprise, uh, ready, scalable solution security, ethical AI practices still need urgent attention from the community. And, you know, it is needless to say that we can foster all this collaboration through these events and, you know, blended with like rapid prototyping.

And, uh, if the open source community comes together, we can play a big role in, you know, developing this ecosystem through these kind of hackathon. So, I mean, so our take on this was like, you know, this capacity, uh, of building new skillset, or, you know, developing a new organization model will not happen, uh, seamlessly without, you know, communities coming together, leaders coming together and ensuring and exploiting this kind of, you know, um, events and trying to kind of understand that, you know, how, uh, we can develop and he, how we can foster this change. So I, I was really excited to be part of this as Canada DevOps community of practice.

I mean, we came together with John Willis and other senior industry leaders from in Ottawa. So there was a lot of, uh, industry attention, um, which was really good to kind of see and, uh, also, uh, explore that, you know, what potential we can, um, you know, these hackathons would carry. I mean, I, I think this is kind of a game changer in a way that, you know, we are, you know, starting to see that it's big events and conferences are kind of translating into micro hackathons, localized, you know, um, hands-on, uh, approaches where people can come together and collaborate.

Well, Great. That's something I wanna talk about. That's something I wanna talk about, because hackathons have taken on, on a whole new form, right?

It used to be get people together, combine 'em together in one location, they go off and work on their own thing. And it's kind of a show rolling show and tell, um, two, maybe we're solving a particular problem, or, you know, in security we do a capture the flag that's essentially hackathons for security, uh, in our environment. I think now be, because, you know, I cover a lot of what, what is the, what is the future role of software engineering, software development, all the roles in the software development life cycle with AI driving or, or part of that picture, whatever it is.

And, uh, where I think we're gonna talk about one of, one of the articles that I wrote on an upcoming SE session of the gang, um, it's come, it's become something of let's get together and talk about what, what is the impact and how would it change not just our individual work, how would it change our DevOps pipelines? How would it change CICD? How would it change not just how we test software, but what, how, how it all gets invoked and how it happens and how we automate these things.

So it's, it's more of a systemic change rather than an individual work change. I'm, I'm curious, Terry, and Rich, as you listen to this, is there other opportunities in the security world to kind of think about beyond capture the flags, types of activities or, or, um, you know, doing red AB type exercises, tab table exercises? There are other opportunities for doing hackathons in security.

Sure, I would think so. And for all the reasons that you stated, I mean, this is sort of a richer environment now, right? These, the, the kind of hackathons that, uh, Gima is, is, uh, is talking about.

Um, I, I also think there's an aspect there that you really are, you know, bringing innovators, students, executives together to really solve, you know, more systemic issues, global, you know, issues that I, I, I believe that there's a place, because security to me sometimes seems so fragmented, you know? Mm-hmm. And, um, I, I do believe that, um, this is, I was excited when I was reading, by the way, uh, Garima, kudos on, on the, the piece that you wrote.

It was so thorough. Um, and I was really excited when I, when I read it, and I was anxious to hear what you had to say, um, uh, this morning. But yeah, I think security could take a, a page right out of this book.

Um, and, you know, I have to, um, add on to, it was great comments, um, Grima and to what Terry said, and I think there's some specifics, you know, for, there's a few of us that may have been around 58 years ago when a movie came out named the Graduate. And there was a line in the graduate that said, there's one word, plastics. If that movie came out today, there would be one acronym, ai, that would be the phrase.

And I think we need to look downstream. It's some of the same ramifications. It changed our lives.

It made a lot of things better, easier, it allowed us to do things we couldn't do before. But now we're also dealing with what happens with all the residual of what occurs when you make and use plastics, right? And, and we're having to deal with some of that now.

And I'm concerned to a degree, and I know this was covered by an earlier game, that there are some long-term effects of using AI and putting AI in place that are both, um, psychological, um, productivity wise and environmentally that we are not yet considering. Now, from a security perspective, without question, um, security is missing from most of AI implementations that I've seen. And the people I've talked to have said, said, well, that's something we can always add on later.

Well, although that's true, you're gonna be less effective and it's gonna cost you at least three times as much to do it. And by the way, in that gap, whatever vulnerability you have is gonna very likely gonna be exploited, because any vulnerability that touches the internet is exploited within seconds today. So I would offer, that's the, that's where the hackathons are gonna help us for security, for ai.

I love that. That's great. That's a great, that's a great an analogy to, to the graduate.

Now, you know what, Jeff, you're, you're onto something because there are a couple of books that just came out, including one called Empire of the ai, which is getting a lot of buzz about the residual effects in third world countries. Like, uh, we, we talk about what's going on in, in countries outside the US in terms of resource use or exploitation of labor that's starting to surface now. So there are longer term consequences.

I just wanted to add that in because that the, the, this book in particular delves into it. Well, I'm gonna have to tie a bow on that 'cause we're, we're outta time on this segment. Yeah.

We'll do a part two and a part three. We're actually kind of gonna roll it a little bit on our next segment. So let's take a quick BA break and we'll be right back.

Discover Textron Group, the epicenter of tech innovation. We are your go-to for reaching IT, leaders and practitioners worldwide. Our secret impactful content that sparks awareness, engagement, and top quality leads with us.

You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more. Join our satisfied clients. Let's revolutionize your tech journey.

Contact us today and tell your story to the world in the most powerful way. With Textron Group. You know, we were just talking about DevOps and, uh, hackathons with ai and how, what's the impact of it, and what kind of things does it allow us to do?

You know, AI observability, et cetera, that were, you know, covered and discussed at the hackathon that REMA was a part of in Canada. I, I think AI is probably even more explosive in the platform engineering space, and we're just kind of getting there. It's just coming to market where we see products that are, uh, releasing not just AI built into the product, but AI in the process and the workflows and things that we have.

Um, there was a, um, an announcement from Stack Gen about their infrastructures code and some of the automations agenda capabilities that they're introducing. And you're seeing more and more of this, um, rather than we're just using AI to do smarter testing or AI to do smarter analysis of what we, what we have, what we know, or getting, you know, information about the code base. I'd love to hear some more thoughts about let's step beyond just the DevOps conversation.

When we talk about platform engineering, that's taken a real systems approach to what we're building, right? It's the foundational elements and what some of the roles, uh, a I AI might play there. Um, John, are you, in your pursuits in talking to companies, what kind of vibes are you picking up?

Speaking of vibe coding, are you picking up? Well done On this topic? It's, it's, um, it's discussed constantly.

I mean, it's the center of the conversation among other things. I mean, these companies are scattered. I mean, they're, they're preoccupied by several different things going on at the same time.

But this is, is definitely a, a main, a main concern, or I would say even an obsession. Um, and, you know, it's funny, uh, with Jeff, I mean, really at home, I, I'm, I'm so glad he, he talked about this that was really brilliantly framed in terms of AI and how its use offers both promise and peril. And I think we're just gonna keep seeing this, this two handed approach.

It's almost like the liberty, the, the, the justice scales one outweighing the other and, and trying to find that fine, fine line between the two sides. And, um, I know, um, I'm gonna point out a few. I'm gonna write a story about Q Herman, a report it did on, uh, CIOs and their attitudes towards AI and infrastructure.

I mean, it's, they're all in, but, um, there are long-term consequences for everything. And I just think it's starting to bear its ugly head. Now, Do John, do you think people are, um, I mean, we talked a lot initially when, especially when Gene AI kind hit the scene about there were no guardrails in place and people were hesitant to move ahead and, um, and rightfully so.

But do you think they've overcome some of their I think they have. Yeah, they have, right? And I think, I mean, and I don't know.

I, I can't, I'm not going to, I'm not, you think That's right though? You think it's a little premature? Yeah.

They got Over it. They looked, they looked at the escalating market analysis and the, and the num the numbers and the estimates, and they thought, s**t, let's, let's just deal with the problem. We, we come across it for now, let's just dive into it and try to take advantage of this.

So there is kind of like your hurdle forward and you pay the consequences, or you, uh, you tack to a different course or you try to correct for the mistakes. Um, and I think you kind of are seeing this with all these acquisitions in cybersecurity. I mean, it's, it's an acknowledgement or a signal that we need a more comprehensive solution.

Uh, ransomware data breaches are seemingly daily, as you know, Terry. Yep. And I think that this is, you know, especially the larger companies that can afford it are trying to put together these comprehensive, uh, solutions.

I mean, Jeff could speak better to this than I can, but that seems to me apparent they're trying to fix things before things unravel, Before they unravel. Uh, I it still feels like, you know, security in particular is, I, I think we mentioned this in the last segment, bolted on, right? It's an afterthought still when it comes to ai.

This Is a theme that, as Mitch can attest, this is something that Alan talks a lot, lot about. And I know you experienced this with him, uh, Mitch, right? This, this, this is kind of, it's the last thing considered, or among the last things considered, but now it's probably, uh, it's more of a priority as we can see from all these deals and, and the, the strategies of the companies.

Yeah. Same conference, new t-shirt. We've been here before.

Yeah. If we've had this problem. Exactly.

So Lemme Lemme throw this out. Um, I, I think, um, I'll, I'll issue this as a po a positive challenge. Uh, no one should wait to get involved in ai, especially security people.

The last thing you should do is wait for somebody to ask what should we do about security? That's almost your fault for letting that happen, right? Right.

You need to be engaging with people, creating software, engaging with the p people who are using agents to automate workflows in the business units. Um, because, you know, it's understanding security means you have to understand what you're trying to secure. So you need to know what people are doing with these tools.

And even though we have great imaginations, we don't think of all the things and all the implications of that. So I think it's, it's a challenge. If you see a hackathon going on, you go to it, it doesn't say security in it.

You go to it, you go to it, the meeting going on about ai, it doesn't say security is here. You go to it. You go and you talk and you learn and you commun and you contribute.

That's how you get a seat at the table. You don't get invited. 'cause you get invited last if you wait And, and then you change your company's or organizations discipline, right?

When you do that, they start thinking of security. They start put introducing security into the conversation a lot earlier, which is super important. And especially when you're talking about something like platform engineering, you know, that should be an early on, you know, security should be there from the get you talk about left, but it should be way over there, you know?

Yeah. If we waited for people to invite DevOps to the table, we'd still be waiting. Right.

Alright. So I've been, I've been doing security for 50 years, and over the time I've been a programmer, I've been a security administrator, I've been a, a chief security officer and, and, and everything between. And it's been about 50 years.

And it's wonderful that the choir still sings to itself, that security needs to move and, and not be the last thing we do. And it amazes me that 50 years later we're saying it a little less often, but we're still saying it. So, you know, I'm not sure we're gonna crack that nut.

You know, Terry, you said that early on there were no guardrails and everyone was hesitant. Now there's no guardrails and everyone's rushing in. Exactly.

I think that's actually the only difference. Exactly. That's what scares me.

I mean, you know, I, I think that's a, a dangerous place to be, but I don't think there's any chance of stopping that. You know, I, I think, and maybe some of these large deals are going to show us something, you know, more comprehensive, that'll, that'll, that'll stop things before a lot of damage is done. But it's certainly not Proactive.

Prefacing, there are two sides. I throw some notes to light to this as well, uh, Jeff, uh, if you allow me. So I mean, again, um, this needs a lot of attention of course.

But we also have to pat the back of the people who are doing great job in this. So I would, uh, reflect MI three has come up with this Atlas, uh, framework, which is great because it gives you guidance on, you know, how to deal with ai, you know, native, uh, software development, how, what kind of practical challenges you can get from a security point of view. Similar framework like Maestro London, all these, uh, you know, frameworks have taken, uh, you know, on the next step and trying to induce like how, what kind of privacy regulations, compliance, security, you know, uh, shift we are seeing from an attack vector perspective.

Right? And another thing which, uh, I think Techstrong, uh, group wrote about this, uh, this week was, uh, this Amazon queue vulnerability, which was found, and, you know, it was how quickly it was discovered and it was no damage done, right? So, I mean, we have to pat the back of the people who are doing this great job.

I mean, I have to kind of say this to all the security professionals. I, I would offer that that is the real good news and that's what should be taking a place in many cases of capture the flag. It's no longer, you know, break in and, and take over it's analyze and find how someone could break in.

And there's a real big difference there. Simply finding an opening is very different from finding all the openings and how we're gonna close them. And that's what should be taking place of, to answer your earlier question of the capture of the flags in many cases.

And Terry, you're absolutely right. Any security professional that said, I've met some that said we don't allow AI at our organization. Oh, really?

Um, how, how can you enforce that? Yeah. Seriously.

Probably going just as well. There's no shadow it. Exactly.

Exactly. So my usual response to that is don't worry. Your successor will take care of it.

Yeah. Maybe sooner than you think. Well, you know, it's, it's interesting you say that, Jeff, about, about the hackathons, you know, maybe the, you know, we take the capture the flag idea and we turn it into, alright, we're looking for the, the Ronan or the Rogue agents.

Let's make it about how do we round up, how do we find these things? How do we know what they're doing? How do we, let's learn, right?

Because part of that is you can't read a book. 'cause you, they're writing the agents right now. So the books haven't been written yet about what these things can and will do and what the telltale signs and how, you know, when they're doing things they shouldn't do, or what security frameworks are being used or not.

You know, to your point, Rima, it, it's, it, we are in the formative time with ai. It is taking shape, you know, and in an ongoing, it's like a river taking shape, right? It's just constantly moving, but it's forming and, uh, that's the time to jump in.

As long as you don't jump in the rapids that are too deep, unless you got a life vest on. But, you know, do that too. You know, get, you know, it gets the, uh, attention of these, especially the companies out here in Silicon Valleys, things like, uh, bank of America came out with a report on the size of the AI agent market in 2025.

They're saying it's 155 billion. I mean, it's just, that gets their attention above all else. Mm-hmm.

And that's just exactly what we do. That that's how they think. A lot of those numbers are driven by the efficiency gains that they think will get.

Oh yeah, of course, of course. Depending getting efficiency Employees. Yeah.

Yeah. Well, yeah. Employees, but also changing nature of people's jobs, Changing nature and things are more efficient, exceptable faster.

Yeah, definitely. There's an upside. Mm-hmm.

Yeah. Definitely is. Well, let's, uh, this is a game we're playing.

We're not even in the fourth quarter yet. I think we're on like, on the second. Yeah.

We're holding Fall beginning In the first, in first quarter, so, you know, keep going. So we'll, we'll take a break here and come back for our next topic. com is the leading resource for news analysis and education on challenges facing the cybersecurity industry.

com covers all aspects of cybersecurity, including data security, DevSecOps, cloud security, application security, network security, security threats, and more. com has the largest selection of security content featuring breaking news, blog posts, podcasts, and more. com to learn more.

com. Home of security bloggers network. We're back on Techstar and Gang.

And Terry wrote a, I believe, a very interesting story about McDonald's and an AI hiring bot named or nicknamed Olivia that was intended to streamline the job application process, but as she wrote instead, it created a supersized privacy headache for McDonald's. Kudos for that lead. And, uh, Terry, maybe you can explain more.

I love the lead. Well, I thinking, I mean, you know, look, the story lent itself to all of the, you know, McNuggets of information. I mean, there was just all sorts of things, ways to go with this.

But yeah, I mean, it, it's, it's interesting. I, you know, McDonald's tried to build a way and in this AI based or bot based, um, recruiter named Olivia to sort through what, you know, must be gazillions, billions and billions served right? Of, uh, of resumes, um, for people who wanna work at, at McDonald's.

And, um, and all well and good, but a couple of researchers were on like, I think a subreddit or something, and they saw that job applicants had been complaining about Olivia and some of the weird kind of answers they were getting from her when, you know, they had questions during the job process, plus it was like a personality test, um, that you had to take to, to get through this whole process. So those guys, of course being who they were just sort of, you know, jumped in and started checking it out. And, um, they ended up finding like sort of, um, the, the company that had that had set up the spot, um, had a, a restaurant owner login.

So if you were a McDonald's restaurant, you could look at all the applicants that that came through, right? But it was very easy to, uh, get the credentials 1, 2, 3, 4, 5, 6. Um, which, you know, just, I dunno that I, anyway, um, so that's what they did.

And then they could see all the information of everybody that applied, um, to McDonald's. They could also see, I mean, they went through and as they enacted, like they were applying, they also found the personality test to be super creepy. And they also found that Olivia wasn't all that helpful on, you know, some things.

But here, all this data is, uh, is now exposed. I mean, to McDonald's credit, they, you know, fixed it sort of right away. But this is just filed under something that never should have happened.

I mean, how many times have we heard about something being left open like this and really poor credentials, like in sort of the development phase, you know? And you know, here you go. If you applied at McDonald's, now everybody knows and they know what you had to say about yourself.

They know a lot of your personal information and there you go. Well, we talked about a lack of security. How about a lack of testing?

Seems like it's kinda some fundamental things might have been missed or under underdressed. You know, it's easy for me to critique, but I didn't also didn't capture 64 million people pray personal information either. So I think I have a right to, to at least be somewhat critical.

You know, it, it's, it's interesting of how much do you create, create something with ai AI and turn loose because there's a lot of value of putting things into the wild out into, you know, systems and applications and whatever you're running to see what people will do with it and will learn. 'cause you know, half of half of planning is executing. You gotta get out there and do it and to see if your plan's even close.

Um, it, it, it, what it reminds me of though, Jeff, is we can't just learn security in the wild alone, right? We've gotta address how we secure these things both upfront and learn what we get when we do put it in the wild. Yeah.

And that's where we apply the lessons of what we learned in the wild prior, you know, there's been a McFlurry of activity around this. Sorry, you didn't use that in the article. I heard that.

I'm gonna say super, you know, can I, my AI Do a follow up now. Alright. Um, but, um, there's really two issues going on here.

One, I believe you had an immature AI application, which really it's immature, it needs to be improved. And then you had horrible security. And those two really aren't directly related, although there's a common core source behind them.

If there had been proper security measures, even industry-wide accepted security standards that said you can't have, you know, sequential digits or maybe even use MFA or something else to be able to get into the system, then all you would've had would, it would be in an immature AI system that really cause has its other own set of issues. But because these two issues are blended and you've disclosed information of 64 million people, the whole thing has a attain to it that's going to be hard to remove. And these are some of the lessons we're gonna have to learn as we go through this.

Right? And this was such an easy one though. I mean this literally the OAS top 10 mm-hmm.

Things to not do, right? I mean, come on. Uh, you know, but you're right, it's an immature system.

It's also, uh, security again is, is an afterthought. You know, I think you I wouldn't even give it that much. Yeah, no, no, no.

I, I don't think they actually thought of it until this happened, and then that was it. But, uh, I also think, you know, and they don't have well-trained people to, you know, spot these things as well. And then, you know, it's just all these topics that we have touched on during this particular text, trunking sort of, uh, apply here, the rush to market probably with this application, you know, before it's well thought out.

And there you go. You Know, Starbucks, Starbucks is another franchise that's going to be trying to u they're trying to be, use AI to help them, uh, uh, um, pinpoint, uh, demand for certain types of drinks or, or to, uh, address fixed machines that are broken, which happens quite a bit at Starbucks. Uh, so I mean, it'd be interesting to see how, how they, how their journey goes with, with McDonald's.

It's, it's weird to me because years and years ago, and again, the CEO of McDonald's has changed multiple times, but I just remember being at a, uh, uh, moderating some, uh, McDonald's CEO chat at South by Southwest and at the time, it's a long time ago that they seem to be on top of things, but it just shows that if you don't have the people in place at the top, anything can happen. Well, they, yeah, that's right. And I mean, I think McDonald's used to be, if not the gold standard, at least somewhere up there, right?

I mean, they were, you know, um, they were a company to look at and being progressive in technology and how it was used and understanding their marketplace and all that. Maybe now that's just a free for all. There's so much competition.

I don't, I don't know. Now I will say about Starbucks, and I find this interesting on a personal account because they got rid of my favorite drink off the menu, uh, in March, the chai cream frappuccino. And it's been like 25 years of ordering that every day.

Now. I can't use the app. I have to go in and have somebody do it for me.

But, so I'm glad I'm gonna maybe start getting people to go order that so the, their system will, will see that that's an important drink to have. But Starbucks is another one though that, um, I don't know exactly where they are now, but I wrote about them years ago. And they're, they're a lot more disciplined when it comes to technology and when it comes to their marketplace and also when it comes to training their employees.

Yeah, they, they, and they, in this case, it was very controlled and narrow in its focus, so it would have minimal damage if something went wrong. And I also think about, was it McDonald's? I know that, uh, Wendy's was trying to drive through AI automated voice to take your order, which was To adherence.

Oh, there's still not, it's just total disconnect S first tion. That's right. So now, now, yeah, the same, the same AI they use on the New York City subway now, I guess, or something that they don't use.

But, um, yeah, so now I guess it's Burger King's turn right to, to step up and show 'em. Oh, geez. Well for that'll be a, Something you that I'm now advocating for.

And that is, we talk about the human in the loop, right? But I think we need more than human in the loop. 'cause that's, there are humans in this loop, right?

That didn't, that didn't cause it. I think with ai, we need to reposition as humans leading the loop, right? It's our, we we are directing what's happening, right?

It isn't just oversight and, and reviewing and seeing if it's okay and accurate or whatever. It's directing where we want it to go, whether we do it or, or AI does it. And ultimately, yeah, that's, we're responsible.

So I think we need to change the model. It's, we lead the loop, not, not ai. That's our role.

You know, I think actually we're kind of, I think I, I sense a transition towards that, Mitch. I really, I really do think that there, as we kind of dive into this stuff headlong and then there are mistakes made maybe that maybe the, that's good that these things are happening now rather than later. And maybe the human has more of an impact, I hope in terms of, uh, work, job experience.

And with the use of ai, maybe the human will lead AI more than the other way around. There you go. Well, we certainly have a lot of, uh, unmet expectations when it comes to ai.

Some, some maybe, uh, a little over exuberant, maybe not. But I, I think our role in it still is a, a very big role in it, so. Great.

Well, thanks guys. Thanks everybody. This has been a lot of fun, Terry, rich Garima John, uh, getting us all together in the gang and talking about, uh, both what's happening now, but probably even more importantly, where we're going and how do we, how do we grapple with some of those things and how do we work together to solve them with AI or without, in both cases.

So we appreciate everybody join us today. We've got a lot of great programming coming up. Uh, John, we're looking forward to your articles this year.

I know you're working on C Out. Thank you. We're gonna talk about this human in, the human in the loop versus leading the loop at, at a one of the upcoming Textron gangs.

So thank you everybody for joining us today. Stay tuned. Be sure and check out all of our other great content on the Textron websites, as well as on Textron tv.

Take care, everybody.