24 Nov 2025 · 10 min
John Jonas explains why he avoids AI in outsourcing and hiring. He uses multiple content detectors to catch AI-written applications, bans generative AI from his team's business content, and insists on direct human communication during interviews.
Welcome to the Secret Sauce of Outsourcing podcast that's dedicated to making you better at outsourcing or insourcing as. That's what we do. This is episode 398, how to handle AI with OFS. So this is going to be a really controversial podcast episode. If you're listening to this, you probably already know that I'm not like super middle of the road on things. I'm kind of extreme. I go one way or the other. I go pretty hard. And we're going to talk about AI, which is very polarizing. People love it or they don't like it. And I am fully, fairly anti AI, actually. So let's, let's kind of get into this. So how do you or I handle AI with your ofs? So here, here are a few thoughts. Number one, when I'm hiring someone, I want to know if they have used AI in the application process. Because when I'm hiring, I want to see them, not GPT. Four, I don't, I don't want to talk with. I want to talk with the person I'm hiring. I don't want to talk with ChatGPT who, yeah, maybe they use ChatGPT later on, but I just let someone go because of it. I hired someone, duped me. They were using ChatGPT to apply, and they used ChatGPT the entire time they were working through me with me. And every single time I communicated with them after the application seems so great. And after that it went downhill. It was like, what? This is weird and this is weird, and this is weird and this is weird. So during the interview process, I want to communicate with them. I want to see your personality, I want to see your English, I want to see how you communicate with me. I don't want to see how ChatGPT does it. And I don't want to see their ability to use AI unless I'm hiring them to use AI, which I'm not. You might. So in my job post or after the job post, when I'm communicating with them, I will either say in the job post, don't use AI in your application, or afterwards. And this is more common, I will ask people, have you used AI during the application process? And typically they're pretty honest with me. I mean, I know they're pretty honest with me because almost every single one of them has said yes, I did. And so, so like the most recent time I asked it, they said yes. I used it in my initial application because I wanted it to take all of the data from my resume and form it into an application. But then everything since then has been me communicating with you, not through ChatGPT. Now, if someone says no, or, I mean, if they say yes, whatever, I'm going to put their. I'm going to put some of their content into a. An AI content detector. Now, if you search AI content detector on Google, you're going to get all kinds of things. And typically what I will do is I'll use four, five, six, seven of them. I'll just copy and paste a paragraph into every one of them and see what it says. And some of them will say 100% AI and another one says 100% human. Right? So those ones I'm gonna disregard because they're not. It's not. It's not reliable, right? Some of them, though, will say, this is 35% AI, this is 85%, 85% chance this is AI. Or this is a 10% chance this is AI. Right? And so by doing that a bunch of times, I get a really good idea if this is AI or not AI. Content detectors are not perfect. Some of it, they can't detect at all. Some of them just say 100% AI every single time I think, which is, obviously you'll learn these, but then also as you do this, you'll start to learn, okay, some of these detectors aren't reliable, but this type of language is not common. This is AI type language. So, and you'll start, you start to recognize what's AI and what's not. If something reads too lofty, it's like, this is. There's too many vocabulary words here that are not normal. Most people have a vocabulary of like a thousand words, and they use a thousand words over and over and over again. And they're not using extravagant words or they're not using extravagant words, like every fourth word. So there are six of them in a sentence, right? And that's where I let someone go recently, because that's how their writing was. And it was finally like, okay, this isn't working. If I find any more AI in anything, I'm gonna fire you. And the next thing they wrote was incoherent, obviously not AI written. And obviously their English wasn't even good enough to write. It did not work for me. I got duped. So that's the next piece of, of how do I deal with AI in my business? I have said specifically to my team, I do not want AI anywhere near our business. I don't want and not AI. I'm okay with Grammarly I'm okay with Grammarly to improve what you have done to make it perfect. I'm not okay with generating content through AI. I don't care how good the AI is, it's fake. It's no human wants to read that when they're looking for an answer. And so like, it may be generative, but it's not truthful, it's not factual, it's not correct. And I don't want it near, near us. And the second reason for that is it's not good for humans. Google says it's fine to use AI, they're lying. Google for sure will be detecting AI and for sure will be punishing things that are just content generated things that aren't human, because things that aren't human, frankly, aren't very helpful. So yeah, you may say, oh, Google said this, that. Yep, Google said that. I've been dealing with outsmarting Google for 20 years. And Google will say all kinds of stuff. So will anybody that is looking to gain attention, they'll say all kinds of stuff. Facebook says all kinds of things that is not correct and true. And Facebook, you may get a bunch of attention on AI post right now on Facebook, but I guarantee that Facebook's gonna be detecting it and punishing it in the future. I don't actually know what's going on on Facebook. I never get on Facebook. I'm just talking about how tech companies work. But I do know that humans don't like AI. And so tech companies, while they say, oh, you're fine to use AI, that's fine, they're not gonna like AI because it's not good for users. So now I know some of you are saying like, but, but, but no, yeah, no, if you're not using AI, then you're irrelevant. Well, I don't agree with that. But I'm also not super visionary. I'm not really good at telling the future. And obviously everyone thinks AI is the future. Everyone also thought that blockchain and crypto was the future, you know, three years ago. And I mean, that kind of didn't really go anywhere. Not that it's not going to be part of our future, but you know, it didn't like take over the world like everyone thought it would. And everyone thinks AI is going to take over the world right now. And I don't really think it's going to. So I'm not saying that AI doesn't have its place, I'm not saying that you shouldn't be using it. But what I am saying is I want to know if something is AI and I use two things right now for that. Content detectors. And my gut feeling of this is weird, which that's why I don't want it anywhere near my business because this is weird. So last thought about this, about how I deal with this, especially in the hiring process because you will find that most people are probably using it in, in the process. At some point you should be asking about it. So I don't do live interviews ever. I never get on live calls with my team after I hire them. I'm not going to do it before I hire them. I can probably count on two hands the number of times since 2005 that I have ever gotten on a phone call with somebody on my team. A lot of you do do live interviews and it is a good way of eliminating a ChatGPT user because they can't use it when they're talking with you. It's also an exceedingly frustrating process to try and get someone from the Philippines on a live interview with you where like you'll have 10 candidates and you'll successfully get two of them onto something and five of them will just disappear when you ask and three of them will schedule with you and not show up. And it's just super frustrating. So I don't ever do a live interview. I've also found that spoken English and written English don't necessarily correlate. In the Philippines, for example, I've seen a number of people whose written English is flawless, but their spoken English is really, really poor. And I know that because I get videos from them and often they're very self conscious about their spoken English because they don't want you to feel like they're letting you down because their written English is so good and their spoken English has such a thick accent that it's hard to understand. I had a programmer once, really, really early on who was an amazing programmer and we never had any problems communicating in written English. I got on the phone with him one time because I felt like, oh, I need this because it's urgent. And I couldn't understand him, not at all. And I never talked on the phone to him again. But writing was really, really easy with him at the same time. At one time I had a girl whose spoken English was flawless, perfect, perfect. Like no accent at all. It was amazing. She had lived in Michigan for a while and she was back in the Philipp and we hired her and I tried to get her to do some writing and it did not happen. Would not happen. She would not write, even though she was, she had perfect English. She wouldn't do it. She hated it. And so I've just found that, like doing a video interview for someone who's going to be writing for you or is going to be doing data entry or someone who's not going to be doing video things for you or spoken things is not helpful in figuring out if this person is a good fit for the job or not. So to me, doing a live interview solves one problem while creating another or while creating multiple others. Like you're going to eliminate a ChatGPT user or you're going to find the best video interviewer, but that does not correlate to how well they're going to do the actual job that you want done outside of not doing the video interview. I mean, I lay out my process for how I find people@1vaoa.com and I don't really have anything outside of that that I do. I know it works. It works really well. Except when I got duped by someone using AI. And that will not happen again because I understand it now. I had ignored it up to that point and now I, now I'm like fully on board with understanding. Like, okay, here's what's going on with AI. And it's pretty easy. Once you do a little bit of, a little bit of reading about it and see some, see some content, it's just weird. And so that's, that's how I deal with AI.
