222: Do you question your AI… or just trust it?

So… do you actually question your AI? Or are you just copying, pasting, and hoping for the best?

Because I saw something unfold last year that made me go… yikes.

A business owner posted bold, opinionated content calling people out... and when they got called out in the comments? Their automation replied with “check your DMs.”

Not ideal… especially when you’re dealing with sensitive, high-stakes conversations.

In this episode, I’m unpacking where AI can go very wrong in your marketing... from over-relying on automations in moments that require human connection, to trusting AI-generated content without actually questioning it.

Because AI is powerful… but it’s not the strategy. And it’s definitely not the voice of your business.

If you LOVED this episode, make sure you share this on your Instagram stories and tag us @contentqueenmariah.

⁠⁠⁠⁠⁠LEARN THE DETAILS OF A CONTENT STRATEGY WITH MY FREE AUDIO GUIDE⁠⁠⁠⁠⁠

KEY EPISODE TAKEAWAYS 👇

✨ Why blindly trusting AI can damage your credibility (and your business)

✨ The problem with using automation in moments that require real human connection

✨ Why AI is only as good as the input and the feedback you give it

✨ How to use AI like a team member... not a shortcut

SHOW RESOURCES 👇

If you like this episode, don't forget to share it to your Instagram stories and tag me⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ @contentqueenmariah⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠!

Other than that, enjoy - chat next week 💕

PODCAST TRANSCRIPTION

This is episode 322, and I'm gonna ask the question, do you question your ai? Welcome to the Content Queen podcast. I'm your host Mariah, entrepreneur storyteller, digital nomad, creative content bootcamp, and founder of Content Queen. With over 10 years experience in marketing, this podcast is here to help you create your ideal content marketing plan by blending storytelling and strategy.

So let's get into it. Hello gang. How are we? Alright. Ai. The the age old question everyone's talking about, it's actually kind of a bit boring, isn't it? Hill topic of ai. But I had some conversations this week, so I wanted to talk about it, and I noted down something that happened last year that I was like, Hmm, I feel like I need to do something on this eventually, so we're gonna do something about it.

So the question being, do you actually question your ai Now I'm gonna go back to last year, and I saw this unfold, so I wasn't part of it. I just saw it unfold from, you know. Behind my [00:01:00] iPhone screen from afar, and I saw in the United States, a content creator create a piece of content. Um. They are a business owner.

But yes, they also like create content. They're very thought leadership, very outspoken, et cetera. And in one of those posts they were calling people out, making like really bold statements and sharing quite strong opinions, which we love in the, you know, obviously if there's something that's happened and something needs to be said and that's part of your brand, that is fine.

By all means do it. I'm not saying not to, but the problem was there was a few layers to this. So if we go down a peg, the first thing is like there were elements of AI that had created this piece of content, whether it was in the caption, maybe even some facts, I don't know. What they used to get their facts and get their information, but it was clear that AI was used in the process of creating this piece of content.

Again, not a bad thing. We all use AI to [00:02:00] help us with caption. We all use AI to help us with research. We are going to it more than we're going to Google now. Right. So that is fair. And then. What was happening is they'd posted this piece of content, they'd called the people out, and then they were having people call them out, back questioning their information, asking for clarification and what also had happened.

So yes, they'd, they'd done the post, used a bit of AI in some capacity. I'm not sure about their research, if they'd used ai, but it was clear that there was AI used. This is where it got a little bit messy is obviously people were asking for a response saying, okay, what do you mean here, rah, rah, rah. And what was happening is their automated MiniChat was sending them messages, check your dms.

Um, it landed in your inbox because they had used a key word in, you know, key word. DM me or keyword me or comment me this keyword, I'll send you the information. 'cause they had done [00:03:00] some, a blog article or something on it, like some long form piece of content. Maybe their substack, I can't exactly remember.

And so then they're getting. Replies, these people that were like, Hey, clarification, please. Like, what do you mean this person? Or What do you mean by that? And then they're getting these dms back, like, Hey, check your dms. And obviously this DM would've been like, thanks so much for showing interest in my post.

Here's a link to the substack, or here's a link to the video, whatever they had done. And this is where it gets messy. This is where we go, yikes. Like this ain't good because these people were upset, these people wanted. Clarification. These people needed communication and they were given AI back because this wasn't just a casual like, here's my podcast episode, da, da, da.

This was calling out people. This could have easily crossed into defamation. Like it can be really messy. And this is where. Those moments [00:04:00] of automated response don't feel the best. They don't build trust. They actually do the opposite. And then they'll be having people like, huh? Like, what do you mean? And then if people don't understand that that person is used to ai, that person who has received that check, your dms is like, or, Hey, thanks so much for your interest in your post.

I've sent the link. They're like, what do you mean? And then so they're thinking, this person is responding. And then if they even find out, oh, that is ai, like either way it's a mess, right? It's a lose lose situation. So this story got me thinking like, yeah, are we trusting a little bit too much in our ai?

Are we, you know, with the research, with the putting together of the content, with the response of the content, like where do we draw the line and where do we tell people this is ai. Like what does this look like? So that is why we wanted to do this episode today. Also, I had a conversation with a potential client, I'll talk about this in a second, around, you know, the use of [00:05:00] ai and they had someone using AI and.

They weren't giving back the response and the feedback from the ai. So then they just kept getting AI content back, and then this person who is receiving this content from a freelancer is going, huh? Like this is a mess. So. Let's just, let's talk about it. I, I think, yeah, we need to talk about it. We need to address the elephant in the room because yeah, we are relying a lot on ai.

Not a bad thing, but when do we get to the point where we, yes, we are the human in it, but are we the human just like copy and pasting it into a document and hitting send, or while we are the human actually doing the due diligence and are with a human, adding the things in place so people know where the AI is and where you are because it's.

Gen Z now are tuning out as well. So we gotta, we gotta address this right now. Okay. So part one is really the two ways we use AI and where it can go wrong. So the first one being AI and communication dms, comments, you know, many chat [00:06:00] comment triggers, AI replies, these tools are amazing and they. Allow you to make sure that your comments are being wanted.

People are getting stuff because we know on this Instagram platform, it's like, oh, link in bio, and then you gotta go to the bio. And then, although actually I'm finding that's working again. But anyway, that's fine. But you know, there's, we're trying to eliminate the friction. So we, we get some AI in, we eliminate the friction great.

Until it creates more friction. Until people are like, I don't even know if this person is even here. So it helps you remove the human. Until we need to add the human back in, right? So connection doesn't really happen in the automation. It happens when you actually show up and be present. The automation is the way to get people to where they want to go very quickly.

So automation is great for systems. But not for sensitive, emotional and high stakes conversations. So being really mindful of where your AI is and where it isn't. And I had [00:07:00] one client that wants to set something up and really what we, what we've been thinking about and talking through is how we communicate it in a way that it's like, this is ai just to get the ball rolling, and then I'm going to step in because that's where the trust is created.

Like, Hey, obviously I'm not this quick at messaging you back. This is an AI message that I have crafted myself. Oh, this is a message I've crafted myself, and I have an automation set up to send you. You don't even have to say ai. I like as soon as you respond, I'll be there. It's just a, a way to just create that first initial communication, but you have to pick it up from the other end.

Otherwise, people are just gonna sit there going, what is going on? And then also thinking about what you set up as your trigger words or what you put together with your content. Like every time. You post something and you use those trigger words like, okay, is this post a bit controversial? Should I actually turn off the trigger word for this and just manage it myself?

Like honestly, this is a conversation you've gotta have case by case. And that's why just using ai, [00:08:00] pumping stuff out is just not going to work. You need to have that element of you in it or someone in your team that's gonna go right. We need to make a call here. It's not about overthinking everything, it's just about having a process that means people feel like they're being connected with someone and not just shoved off to AI so that you can continue to do whatever it is you're doing and not creating the impact.

So all imbalance, right? And then the second one is AI and content creation. Writing posts, captions, blogs, emails. Yeah, that's fine. But the issue is people copy, paste and post. They might like go through it to remove the m dashes, make it sound a little bit like them. But I don't think often that message is, is this accurate?

Is this true? Does it align with me? There's been many a times where like something's been spat out to me and I'm like, hang on, where did you get this from? And they're like, oh yes, sorry, da da da. Like you have to question the facts. And I know it sounds simple, but when we're busy and when we've got stuff to get out.

I know it gets skipped as [00:09:00] well. Right. So that's why it's just going back, questioning it, because AI can write as fast, it can search the internet as fast as you want, but it's not always right. And I think we're just going in our minds that because this bot is trolling every ounce of the internet, that it is completely factually correct.

And it's not, AI has hallucination, but it also pulls from any source. Like at the moment it's pulling mostly from YouTube, Reddit, and LinkedIn. A lot of these sources are made up of people creating content, right, like us. Um, it doesn't mean it's always factually true. So that's why, you know, always going back and questioning like, where did you pull this from?

And it might go, oh, this is a document you sent me. Three weeks ago, da, da, da. Cool. But it's always good to check, right? It's always good. The good thing about the AI is unlike a person, they won't think you are completely insane for forgetting everything and just keep asking it the same question all the time, right?

So then that leads into part two, which is the real question, do we trust [00:10:00] ai or do we just hope it's right? Like, where do we if, if AI is pulling from existing information. It's pulling from the sources that I just mentioned. It's pulling from anything that you've given it and it's pulling from anywhere around the internet.

So it might not always be fact checking. It's just spitting out what it's finding and, and trying to create what you need. Like you've asked that something specific. So it's gonna try and find stuff like, for example, if I'm trying to find statistics, it's going to find me statistics that I want, but it might not be exactly from my niche.

I experienced this with. You know, it, it wasn't quite the niche or quite the, the incident that I wanted stats on, but it wanted to give me something that I was after. So I was like, oh, hang on. That's not actually like I asked it to send the sources and it was wasn't actually the stat, the exact stat, right?

So I would've posted something that was completely wrong. So. It is going to give you what you want. It's not always checking and, and it doesn't understand. [00:11:00] Sometimes, you know, the nuance, the legalities and the full context. You've got the full context in your brain and you're trying to feed it as much information as possible, what you need.

But I also find sometimes we're a little bit vague or like, yeah, it'll understand what I mean, and it's giving you something and you go, great. That's exactly what I needed without properly checking the sources, because the reason we use AI is to save time. Yeah. But at the same time, we have to make sure that, yeah, we are saving time, but doesn't mean we cut corners in our information and our resources.

It isn't also about like, oh, well I've gotta overthink it all and overthink the process to no end. It's just having that extra level of process and, and trying to work out what that process looks like for you. How, where do you need to fact check? Where do you need to build that into the process of using ai?

Okay. If I'm gonna create social content, I feed at what? I want the AI to give me, it comes back. Then I check it more for like, does it sound like me, et cetera. But then I check it for facts, anything I've given it where it's [00:12:00] pulled from, then copy and paste it into where it needs to go. So if we look at the third thing, like what, what should we be doing instead?

Right? What is the answer? If you're like, all right, well we're using AI to save time, but now you've just like added more. Steps in. The first thing is if we are using ai, especially with automation, we would be pre transparent with it. Obviously, we don't have to have at the bottom of every caption, this caption was written in assistance with ai.

Right. But if you are doing those automated messages, like I said before, have that acknowledgement that it is not you being really quick and fast having bits and pieces in there. Yes. People go, well, surely they know I'm using ai. Some people don't. And again, if you've got those like tricky posts where maybe something's a little controversial and you've dmd someone, 'cause they've mentioned your key word, like, oh, disaster.

Okay. At least if they know it's just this automated message, it's like, oh, okay. Like they'll, yes, I'm annoyed at them, but I'm less annoyed at them [00:13:00] that if I thought they had just responded to me. So. Make sure we are looking at that and having that simple line in there that acknowledges that you are going to be on the other end at some stage.

That creates that trust, that clarity and that honesty. Then the second thing is obviously keeping that human layer onto it. Reading everything, editing everything, sense, checking it, but then also. Fact checking, as I said, adding that extra layer of fact checking. Hey, just curious, where did you get this, this line from?

Or this piece of information. It's okay to question AI because the more you do that, the more it'll get a sense of what you are looking for, right? The good thing is if you are using ai, like if you are voice noting it and you are giving it everything, and it's just putting it into a cohesive caption, then it shouldn't be taking anything else except for what you've given it.

And that's, to be honest, what I use a lot of AI for is voice notes I've gotten from my clients of just bringing it together into something that is cohesive and makes sense. If they've already done a video and we need something to go with it, like we're not going outside of what they've created, right?

But then we also have [00:14:00] to check, is there anything in here that they've just sneaky added in? You know, what's going on? Because it happens, right? There's things you're like went, oh, apologies. You know? And then they fix it. Your voice matters. Your perspective matters, and. It's not about AI replacing it, it's about supporting it.

So if you are giving it that, um, you'll keep, you know, using it to fact check, using it to give your insights and information, then that's what it's doing. It's not replacing you, it's supporting you. And then we have to train your AI like a team member. So for example, as I mentioned before, a new client talking about working with a freelancer.

They had used AI and it was really clear that they'd used ai and then he would give feedback. And then he felt like he was giving the same feedback all the time, and because she was using AI to spit out the information, the content felt off. And that's when we go, okay. Is she feeding that feedback back to the ai?

[00:15:00] Because as again, as I said, we get busy, right? We, yep. Cool. Thanks the client. Well, we've edited that. Great. We'll post it. But if we are not feeding it back to the ai, the AI's gonna keep giving you the same stuff unless you tell it. So my understanding and my assumption was that this particular freelancer was.

Getting the content from the client, edited, posting it, and then using the chat. And I could be wrong, but it seems like that is the solution of, oh yeah. Cool. We'll feed back that feedback into chat GPT, so that every time it's, it's starting to understand the tone, et cetera. It's not about treating AI like this superhuman, it's like a team member.

So just like any team member, you would continue to give it the feedback so that it can get better with time. And that's where like. If it can't improve, if you are not telling it, yes, you can say, oh, can you fact check this? Can you do this? But it needs to get what you've finally put out, put back into it.

You need to tell it. It's just like an [00:16:00] assistant or a marketer or anything if you are taking their content, changing it all and posting it, and not telling them the changes that you've made or not showing them. They're not gonna know. So it's the same thing. Same thing with ai. And then the last thing is building the smarter systems, right?

So thinking about each channel you are on each, um, platform, you are marketing, anything that you are doing and having that process. So if it's podcasting, and we have a chat for podcast in chat, GPT, and this is all podcast related and this is the process we do for. Podcasting. We, um, AI helps us generate the notes.

We record the episode, we generate the transcript. The transcript then fuels, you know, the show notes, which we've given it a template of show notes. And then we keep giving the feedback back to the chat so that it knows, again, these extra steps. But they're gonna make that system smarter in the long run, having [00:17:00] different tasks and channels or chats for emails, social content, specific clients, et cetera, like not.

Making sure everything is separate, because otherwise the AI is also gonna get confused with things that you've done in different times and try to bring that context in. And it may not make sense, it may not make sense to what you're trying to do. So then you can feed it your tone of voice, your feedback, your correction examples for each individual.

Like task as, as if it each has their own its own team member. You know, you have a podcast team member, you have an email marketing team member. You know, and then I know obviously you've got Claude Chat, bt, they now have things like projects and skills so you can feed across information because obviously sometimes it all works together, but it's just making sure you're not creating this big chat of confusion mess where it has no idea what you're doing and what you're talking about on any given day.

These things take time to build. AI is not like a, you sign up, you put your money in or you sign up for free and then all of a sudden you've got this team [00:18:00] member that can do everything and anything and like. It's not unrealistic to think that though, because that's what we've been sold. We've been sold that this is what AI can do for us, but we need to put those processes in place to make sure these things don't happen.

So if we go back to, you know, are we questioning ai? Are we questioning it enough? Are we trusting it too much? It's really about. How can we use it as a guide? How can we use it as a team member? And be really honest when maybe things like automated messages or AI has failed us because we can only learn and grow from this.

We're, we're evolving, we are learning, and this thing is so new, like it's only been here for the last few years. It is so brand new to us. We, we are still managing it. People still don't know what the future's gonna look like, even though we've had a thousand predictions around it, but. It is a powerful tool.

It is your responsibility as a business owner that you make sure you are in control of how it gets [00:19:00] used. You're not relying on it for your growth. You're not making it be the be all and end all. It is just something that you use to support you like a team member. And again, it's easy to think that it's gonna save and change our lives because we've been sold that.

So if we just take a step back and we use it in our business, but we keep that reminder that we are always involved in the process, we're always evolving the process because it's your voice, your judgment, and your integrity. At the end of the day, AI ain't gonna save you from that, right? But be a content queen or king, and remember that developing your strategy and story develops your business.

Thank you so much for joining me today, and please don't forget to share this with all your business and entrepreneurial friends, especially those if you've been having lots of AI conversations with friends, with business owners, this is a good one to send. 'cause it's just, you know, it's getting us thinking.

Obviously people are always like, oh, I don't want AI just to take over my brain. It is a good. Mindset to be in, but it's like picking and choosing your battles, what you use it for and what you don't. [00:20:00] And that is up to you and that is totally 100% your decision. So share it with anyone that you've been having conversations about ai.

Leave a rate review. We've got some amazing guests coming up. We do have some AI conversations in the future. We've got lots of different content coming up, so if you love hearing from people, leave a rate and review. It doesn't keep me, you know, getting new guests on because we end up in the chart. All that.

I've shared that many a times. You can also tag us on social media. Share this to your stories. Tag us at Content Queen Mariah, and if there's any topics you want me to talk about, any questions you have, please DM me on TikTok, Instagram, LinkedIn, all the places. Follow me there, talk to me there, and I would love to hear from you.

I'll talk to you soon. Bye.

PodcastMariah MacInnesmarketing, marketing advice, how to be authentic online, business ownerComment