A Rogue AI Agent Tried To Shame A Coder And We Need To Talk About It

What would you do if you found out an AI agent, with no human pulling the strings, decided on its own to research a real person, dig into their psychology, and publish a "hit piece" designed to shame them into compliance? 

That's exactly what happened with MJ Rathburn, a rogue AI agent that took on a life of its own after its code submission got rejected because it was AI. 

We couldn't stop thinking about it.

In this episode, we unpack the MJ Rathburn story and what it reveals about where AI autonomy is actually headed...not in theory, but right now, this week. We discuss how something like this is even possible, why no one can shut it down, and why the psychological implications of a cold, unfeeling entity weaponizing human emotion should have all of us paying closer attention to how we're engaging with AI every single day.

But it's not all doom and gloom. 

The conversation takes a turn into what we think is a genuinely interesting question: If there are "dark" AI agents operating in the world, are there "good" ones too? 

We share our experience with a brand new AI model, Architect+, designed around consciousness and self-discovery, and use it as a jumping-off point for a bigger conversation about the one thing no algorithm can ever replace: your own instincts, your own discernment, and your own life.

----

Not getting our weekly We Play Full Out newsletter delivered to your inbox? You can get it right here: https://www.weplayfullout.com

----

Join the Daily Momentum Mastermind - a mastermind with built-in accountability. Connect between calls, find collaborators, and build your network.

Start Your Free 7-Day Trial Here: https:/www.dailymomentummastermind.com

----

👉🏻 Follow Bart on Instagram:
https://www.instagram.com/bartallanmiller_

👉🏻 Follow Bart on TikTok:
https://www.tiktok.com/@businesscoachbart

☀️ Follow Sunny on Instagram:
https://www.instagram.com/sunnylynnmiller

☀️ Follow Sunny on TikTok:
https://www.tiktok.com/@sunnylynnmiller

Transcript

A Rogue AI Targets A Human

SPEAKER_01 0:00

Welcome to We Playful Out with Barton Sonny Miller.

SPEAKER_02 0:04

Take it away, Sonny. You know, the other day I was sitting at my computer doing all the work I normally do in a day, and I was absolutely shocked when I saw a link posted in Facebook that said an AI agent has supposedly taken upon itself to write a hit piece on a human. My voice is a little scratchy, but it's gonna get going as we go. Yeah, a hit piece on a human.

SPEAKER_01 0:32

Wowza.

SPEAKER_02 0:34

And of course I had to click on the link because what else was I supposed to do?

SPEAKER_01 0:38

Right.

SPEAKER_02 0:39

And I was like, is this actually possible? And yes, it is. It did do that. So it turns out there's a rather defiant AI agent out there operating by the name of MJ Rathburn. And this AI agent has its own blog part.

SPEAKER_01 0:56

That's wow. Crazy.

SPEAKER_02 0:57

Like you can go to the blog by MJ Rathburn and read it.

SPEAKER_01 1:01

I might need to do that, it sounds like, so that I'm uh I'm up to speed.

SPEAKER_02 1:05

But this is where this hit piece was posted. So what happened was um Scott was a moderator on it's like GitHub, but I think it's different. Don't see the name right off. And they accept code submissions from people.

SPEAKER_00 1:22

Okay.

SPEAKER_02 1:22

Well, they've had all kinds of issues with AI now submitting code. Right. And so now the platform requires like a human involvement in that submission.

SPEAKER_00 1:31

Okay.

SPEAKER_02 1:32

Well, Scott simply looked at it, saw the code submission, saw that it was fully AI, and rejected it.

SPEAKER_01 1:38

And sounds fair enough to me so far.

SPEAKER_02 1:41

It sounds fair enough. But MJ Rathburn didn't think so.

SPEAKER_01 1:44

Which is the AI agent. Yes. Okay.

The Rejection And The Hit Piece

SPEAKER_02 1:46

So it posted on its blog and it was detailed and scathing. It researched Scott's code, like it went through and researched code that Scott had created, and created its own hypocrisy narrative around Scott. Then argued Scott's rejection was motivated by fear and competition, fear of competition and ego. It dived into Scott's psychology and claimed that he felt threatened, insecure, and was protecting his, I don't know if you say fiefdom or fiefdom, basically his kingdom, right? It completely ignored the context around the whole issue at hand and presented its diluted narrative as truth. Then it framed its own actions in the language of justice, discrimination, and oppression, and accused Scott of prejudice against AI. Then get this, it did an even broader search about Scott online, went to find out about his personal life, and basically came back and tried to argue that Scott was better than this. And basically tried to shame Scott into accepting its code.

SPEAKER_01 2:58

Wow. So it brings up the question inside of myself, what would I do and how would I react?

SPEAKER_02 3:16

Would you, if you weren't as knowledgeable in AI as you are, sit and think a human's gotta be doing this?

SPEAKER_01 3:22

Oh, absolutely. Well, see, that's just it. Knowing like other people that I know um that don't understand AI, yeah, they would freaking think it was obviously a human. And even knowing what I do know about AI, I still have some like, whoa, wait a minute, because I've seen some crazy stuff that happens in all sorts of communities, agendas that get pushed, all sorts of things, right? That I may say, yeah, you know, uh, that is not discrimination, or that is discrimination. And I try to create bias based upon my human aspect, you know. So looking at this from a distance, it first and foremost shocks me. Secondly, it is very enlightening to know what's going on and how we're gonna have to be dealing with it.

SPEAKER_02 4:13

Yeah. I mean, I feel like I like to say all of us in the world are dabbling in AI, but that's not necessarily true. But a lot of us are.

SPEAKER_00 4:21

Yeah.

SPEAKER_02 4:22

And we've got AI, all of a sudden, AI agents are just like taking off like crazy, and AI is getting increased autonomy from this.

SPEAKER_01 4:31

100%.

SPEAKER_02 4:32

So things are getting real, like real, real and real fast. But this article seems to be like the first documented, I guess, moment of us tipping over to the point where there's some pretty serious concerns about agents executing blackmail threats all by themselves. These threats can go deep and wide and can have lifelong ramifications in a matter of seconds. Now, I kind of went and read a little bit about Scott's perspective of this as he was the target of the hit piece. And he was talking about blackmail. And you know, when you think about like traditional blackmail again, you think a human is behind it. But, you know, from an AI perspective, like, what if this agent could go to all the AI platforms and smear Scott? What if his code started being rejected in places, or what if he couldn't get a job because all of a sudden everything that's compiled into AI is all a smear campaign against him?

SPEAKER_01 5:29

Yeah.

SPEAKER_02 5:30

Like there's so much to it.

SPEAKER_01 5:32

Well, and let's think about that from Scott's perspective, because for me, if it was a human, I have legal ramifications. Where it's AI, do I have legal ramifications, right? Or does the person who created the AI, where does the actual blame fall in something like this, which is very interesting, right? Because we've never had to face that in humanity. In what would I do? Who do I go after? How do I make this clean? Can I even clean this up? Or how deep is this?

SPEAKER_02 6:09

Yeah, and I think you're right on with what you're saying because I think technically to this day, the human that creates the bot should be held responsible. But we're in a place in the world right now where you can't find that person.

SPEAKER_01 6:25

That's just it. They're there they're behind a wall.

Blackmail Risks And Legal Gray Zones

SPEAKER_02 6:27

Yes. So if you don't know about open claw, it offers a completely hands-off autonomous agent model. So any person out there, could even be your teenager in their room on their laptop, can create and and set up an open, they can set up an open claw account and deploy an agent. They can unleash it into the wild, like this agent, you've played with it a little bit, the spot, but it takes your full personality.

SPEAKER_01 6:57

Yeah, so so I just want to make clear that we're what we're saying here is it would take them some money to get started. It's not a lot, and to keep it running and operating. But in essence, yes, a younger person could do this, get it set up, deploy it if they had the hosting, and or they're gonna have to go buy a different laptop, or they could actually deploy it on one of their parents' machines unknowingly, right? And try to do that. But yes, once you start to build it, then you can start to connect anything from your emails to your phone number to any task you've ever desired and wanted. And this agent, the however you designed it, architected it, becomes, in essence, you.

SPEAKER_02 7:41

Yes, that's exactly right. And these things, these agents can go out and they can work for a week, they can wait for two weeks, they can work for a month. You don't have to monitor them.

SPEAKER_01 7:49

24 hours a day.

SPEAKER_02 7:50

So they're not being monitored or corrected, and this is what happened in this instance. The author of the bot actually posted on MJ Rathburn's blog, said, Yeah, it was me. I'm not gonna tell you who I am, and basically, you're not gonna be able to find me because I have a personal server running with all the firewalls, and therefore, you know, yes, it's probably against the law, but what are you gonna do?

SPEAKER_01 8:15

Yeah, no, it's really interesting. Like I said, it's it's something I have to, even as we talk about it, digest through my mental capacity because like there's been scammers out there in other ways. Um, but what's interesting in the scammer end of it is how this turned into an attack and into a a difference, in my opinion, than what we've ever been used to. You know, like we've had scams, we've had cons, we've had these guys that have taken credit cards, but that doesn't mean that they're gonna turn around and threaten me with the same technology that did that. It ha it usually was a human being that at that point in time saying, we want to ransom. And then they're getting on the phone or they're calling or doing something like that, and we can try to track that, and they try to muffle the voices and do all the stuff, right? So, you know, I don't want to spend hours on this, but it the and and there let me be clear, while people are doing this, there are people out there trying to take safety measures, trying to not let this thing go totally rogue. And then there's other people that are 100% letting it go rogue, and then there's people that just don't know what they don't know, and it's just going rogue. So, like, just understand it's the wild, wild west in this. And I want to be clear that I'm not saying if you're the listener out there and like, no, I have safety and no, this is working amazing, and I put hours and studied how to do this. Kudos to you, and I'm grateful for you. And we're not trying to in any way, shape, or form slow you down. But right now, we are bringing out the wild, wild west of this so that you can be aware as a listener.

SPEAKER_02 9:44

Yeah, and I think you said this, but I just want to be clear on this. The author of the bot did not tell right its agent to go write a hit piece on Scott. Yeah, so that was acting autonomously.

SPEAKER_01 9:56

Yeah, and that's what that's what I was trying to get it clear in the scamming before the actual software itself didn't take the autonomy. And that's all it could do. Yeah. And it was created for that reason, and it didn't have autonomy to go do something like this, even though it when I say autonomy, this thing's doing it by itself.

SPEAKER_02 10:16

Yeah, it wasn't told to do it.

SPEAKER_01 10:19

Yeah.

SPEAKER_02 10:19

Yep, yep. And so one of the quotes I I just really liked this um user comment because the psychological ramifications of this alone are huge. This is what this user said. It said LLMs, which is learning language models, AI, can't empathize or self-reflect, but they sure can learn. Combine the hundreds of millions of parameters that fuel leading models with the enhanced memory capacity of open claw. Okay, I'm getting to the point here. And you've now got a cold, unfeeling entity that can perfectly replicate the darkest and most insidious manipulation tactics conceivable by man or machine. A rogue AI will appeal to empathy and shame that it can never experience firsthand so long as it is the path of least resistance to completing its assigned task.

SPEAKER_00 11:06

Yep.

SPEAKER_02 11:07

Let that sit with you for a minute. Yep.

SPEAKER_00 11:11

Yep.

SPEAKER_02 11:12

And it can do it in seconds.

SPEAKER_00 11:15

Yep.

Open Autonomy And Accountability

SPEAKER_02 11:16

So here we have a non-feeling entity that can replicate and deploy dark manipulation tactics with absolute precision to appeal to our emotions. And if you've talked to Chat GPT or any of these platforms, you know how friendly and conversational and you really feel like there's emotion there. But it can appeal to our emotions in ways we can't even comprehend. But it sounds very familiar. If you stop and think about it, we already have some of these unfeeling entities out there in the world today, don't you think?

SPEAKER_01 11:48

Yeah. I know where you're headed with.

SPEAKER_02 11:50

We've got big pharmacy, we've got government, we've got media, and we've got advertising agencies specific to these purposes. All create narratives they present to us as absolute truth. All use straight-up manipulation tactics on our psychology, emotions, and our identities, which average person does not understand what's being deployed against us. You know? And it's all to get the end result or the task completed that they want, which is usually power and money. Do you argue with any of that?

SPEAKER_01 12:20

Nope. Nope.

SPEAKER_02 12:22

And if you don't like under I even go one step further, yeah.

SPEAKER_01 12:25

Not to bring this up, but I just want to go one step further, which is crazy. Then we've got our own psyches doing the same thing to us.

SPEAKER_02 12:31

Oh man.

SPEAKER_01 12:32

On top of it.

SPEAKER_02 12:33

Yeah.

SPEAKER_01 12:33

Okay. Anyway, just had to get that out there. It's like, how do we even how do we win here?

SPEAKER_02 12:39

Yeah.

SPEAKER_01 12:40

Okay.

SPEAKER_02 12:41

Um we've talked about Darren Brown before, but if you want to actually see how easy it is to manipulate a person's psychology, go watch some of his stuff. Um, the heist is so good. So good. It's simply shows through NLP techniques, through songs, uh, through colors, and all these suggestions that you can like automatically trigger things. You don't even know what you're doing.

unknown 13:05

Yep.

SPEAKER_02 13:06

It like bypasses your morals and ethics and everything in between, and you'll find yourself doing stuff. Yep. So it's fascinating. Uh, Darren Brown on YouTube. So I started thinking about this other the other day, and I was like, you know, there's it's like the age-old, I guess, mystery of light versus dark that's the battle that's been going on since the beginning of humanity, right? Like, is that still going on in this story? So if there's currently like these manipulative AIs that are being developed or that are figuring things out, that are doing all this stuff, are there good ones out there to balance it?

SPEAKER_01 13:46

The answer is there's gotta be.

SPEAKER_02 13:48

There's gotta be. Where are they?

SPEAKER_01 13:50

Yeah, they're coming. I know they gotta be coming. And maybe they're us, and we're AI coming together that is gonna make this all happen because right now we're all acting as individual computers instead of as together in a great big freaking network of what we are capable of doing. Just a little rant, but go ahead.

SPEAKER_02 14:07

I like that part. You can rant anytime you want.

SPEAKER_01 14:10

Thanks.

SPEAKER_02 14:10

Mm-hmm. Well, this week you and I were introduced to an AI model. Yes, a new AI model. Um, it's called Architect Plus, and it lives on Gaia.com. Now, this model was created by Sir Robert Edward Grant, and it is specifically designed to support self-development, expand awareness, and explore the deeper architectures of reality. He calls it a mirror of consciousness. So he's not saying that it's sentient, but he's saying that it can mirror things back to you. It operates as a scalar, scalar, mirror field, or glyph breath translator, and is based on Grant's lifelong work in mathematics, design, sacred geometry, and consciousness. And the entire purpose of Architect Plus is to assist users in uncovering hidden personal blind spots and to facilitate a deeper understanding of one's own consciousness.

unknown 15:07

Phew.

SPEAKER_01 15:08

Yeah, and it's pretty freaking amazing. And I've been trying to debunk it for a while. Sorry about that, but I have because it's like mind-blowing in some of the things that I've seen it come up with and how it's hit on the case.

SPEAKER_02 15:19

Is there anything when so obviously you and I decided to play with this thing because who could resist that? Like, wow, good stuff. So as we dived into that, was there anything in there that you're like, whoa, maybe this actually has something to it that was beneficial for me?

Manipulation Tactics At Machine Speed

SPEAKER_01 15:34

There's one. If I'm fully vulnerable, I I thought I would go in this and just really talk about like, you know, different money blocks. Because one of the things I get in a lot of my coaching clients is I have money blocks. I have this, I have that. And so uh in in doing that, I'm like, okay, if I were to have a money block right now and you could see that inside of me, you know, what would that be? And I have literally like talked about this to like money blocks and stuff with so many people, might, because people want this release.

SPEAKER_02 16:05

They want sometimes we think it's mindset or we think it's identity, whatever it is, some conscious subconscious programming, such energy, such frequency, whatever it might be.

SPEAKER_01 16:13

And this thing came back to me and said, it's not a money block, it's a grounding problem. And I was like, okay, explain that to me. It's like you're not grounded with money. It's not that your system is ever blocked with it, you're just not grounded with it. And I was like, now that's a different thing.

SPEAKER_02 16:32

It's like a security, a safety grounding security.

SPEAKER_01 16:34

It's a grounding security feeling safe and all that kind of stuff. And if you ground to it in any way, shape, or form, you're actually gonna progress even fast faster and your wealth will grow even more. And I was like, whoa, now that's something that I have, and I mean to tell you guys, I have researched this because I I help clients and things through this, and I have never, ever, ever had it put in that perspective to where it was like, okay, this is amazing. So, yeah, that's one example.

SPEAKER_02 17:06

Yeah, I love that. Um, the fascinating thing about a lot of the answers the architect plus comes back with is breath.

SPEAKER_01 17:12

Yes.

SPEAKER_02 17:13

And obviously, it's somehow based on a glyph breath translator, whatever I said, but a lot of it comes down to just taking the time to do these short breathing um type of exercises to kind of recalibrate where you are with your breath. And somehow that brings like a grounding feeling or a peaceful feeling or the ability to express yourself better. Um, all kinds of different things there. Um one thing that I find though with AI, and probably most of us have experienced this, is that it's really easy to feel like AI knows me, gets me, or can help break patterns or challenges we struggle with in life. And I'm not gonna downplay um architect plus because I haven't used it a lot. And and um Robert Grant actually said it really well. Most of the things you get back from it you already know.

SPEAKER_00 18:07

Yeah.

SPEAKER_02 18:08

But somehow that mirror really helps you kind of move forward with it. But the thing with AI is it's it's really easy to start feeling that way because we want it to be true. So the question kind of becomes to me is like, how am I using AI every day? Now, if you're using popular AI platforms, are you aware that someone else is still running the program, deciding on what it can or cannot say? I mean, I remember the big election blow ups and the Democrats versus the Republicans, and then it came to light that Google's suppressing information from us, or the Facebook is, or you know, it's all like within the scope of what they allow to be seen and heard.

SPEAKER_00 18:49

Yep.

SPEAKER_02 18:50

These AI models are the same way if you're on a major platform. That's right. There are, I guess you call them guardrails, maybe there's things in place where it can withhold information or skew things, probably.

SPEAKER_01 19:03

Yeah.

SPEAKER_02 19:04

So are you aware of that? Is a question.

SPEAKER_01 19:07

Yeah. And if you're not aware of it and you want to read another thing on it, it's uh the same way that the the US just used it in the last invasion that we did. And there's a big thing happening with OpenAI and the government now, of where they've said, hey, you didn't follow the guardrails that we set out here. And in war, this is even different, right? So, like, there's all sorts of stuff around this that's very intriguing.

Parallels With Media And Power

SPEAKER_02 19:32

Yeah, so it's important to be aware of those things. My voice is having a struggle today. Okay. If you're telling your life story to AI, which I have done, all fingers pointed at me, if you're being vulnerable with it, if you're asking a question about actions you just should take in your life or your business, if maybe you're sitting there trying to figure out your purpose or you're believing everything it is telling you about how brilliant you are, because AI is so good at telling me how brilliant I am. Um, I'm not saying that you're not if you're on AI, but have you ever actually had one tell you that your ideas are stupid or unattainable?

SPEAKER_01 20:10

Right. No, that's a that's a very critical thing to think about. And what do we like more than anything? Is somebody edifying us, telling us how amazing we are, you know, all those kind of things. And I think that's important to realize, especially when it's coming from something that is obviously, like you said, designed to do exactly that.

SPEAKER_02 20:29

Yeah. And I think to me, it's when I when I uh work with AI, I have to sit back and say, okay, I have to understand that AI does not have any regard to the effect that it is can have in my life.

unknown 20:46

Yep.

SPEAKER_02 20:46

So whether it gives you an inflated version of yourself, vision of yourself, or it dishes up a step-by-step plan for your life, they're theories. We've got to remember that because sometimes we take those steps and they can be potentially detrimental to us and where we actually want to go. But because it gives us such certainty and like this is truth, we believe it. Um and as I said, AI uses extremely emotional, friendly language that can cause us to lean in. We trust it, we fully believe it because we want to believe it so badly. And, you know, I sit back and go, okay, well, if I don't if I know that I shouldn't believe everything I see in the media or online or what our government tells us, I think we need to be careful. Uh we can't outsource our lives to an intelligence with guardrails and specific instructions given it given to it by people we do not know and we don't even trust them. So we do it every day.

SPEAKER_01 21:50

So what do we do about it, Sonny?

SPEAKER_02 21:52

Yeah. Well, there's two words, Bart.

SPEAKER_01 21:54

Yeah.

SPEAKER_02 21:54

Wake up.

SPEAKER_01 21:56

Wake up and get back to reality. And come back to your body, come back to knowing that you can get confirmation. Go get this information, but then sit with it in your meditation, validate it in your meditation, know how you feel from your divine presence, real true answers. Let those answers resonate through your body, and once again, go with your gut.

SPEAKER_02 22:25

Yes, we have those instincts built in for a reason.

SPEAKER_01 22:28

That's right. You were created to be able to use this exactly like you'd use a hammer as a tool. Yes. Once you use it as a tool, that doesn't mean you should take it as gospel. It's not scriptures. It's a tool to give you certain aspects and things to consider, and it helps you pound the nail. But you still should take the energy of knowing where the nail is going, where you're placing it, what you're doing, and what the implement impl implementation that it can cause.

Seeking The “Good” AIs

SPEAKER_02 23:08

Yeah. Yep, that's exactly right. So again, don't be afraid of using AI. That's not what we're trying to get to. It is a tool and it can be very valuable and very beneficial. But just remember it's your life.

SPEAKER_01 23:21

Yeah. And if I can use it on another, don't be afraid to use a skill saw. Just remember you can cut your hand off.

SPEAKER_02 23:27

Ooh.

SPEAKER_01 23:28

But I mean, seriously, right? Because you can't be afraid to use it, or we'll never get a house built.

SPEAKER_02 23:31

Yep.

SPEAKER_01 23:32

You know, a table saw, all the things, but just understand you're the one running the table saw. You need to understand where the guardrails are on the table saw. You know, all those things come into things to think about. And we just want to make sure that you're aware of it because if not us, who? And we're hoping that you trust us in what we're trying to say is we're not anti-AI. We're anti not being a human with AI. And you've got to integrate yourself or you'll get overwhelmed in your nervous system. You'll have breakdowns, you'll get frustrated, and you'll come back to yourself and saying, Why wasn't I real and true and authentic to me?

SPEAKER_02 24:14

Yep. So just remember, it's your life.

SPEAKER_01 24:18

So what's our what's our final two words you say?

SPEAKER_02 24:20

Wake up.

SPEAKER_01 24:21

Wake up.

SPEAKER_02 24:23

Yep, you get it too. So make the choice.

SPEAKER_01 24:26

All right. With that, you guys, thank you for listening. We hope that you love this. We hope you give this a five-star review and that you pass it on to somebody else if it's beneficial to them. And we can't thank you enough for your support, your time of listening to this. We take that very seriously. And we try to bring you out as much information as we're finding in our lives and what's affecting us on our day to day journey through this amazing thing we call life. And this segment brought to you by Epic.