AI this, AI that. Everybody keeps telling you to use AI, but what if I told you there was a better way to use it?
AI is causing a lot of problems when it comes to social media right now. One of the first things you may have heard about is LinkedIn recently adding an “AI slop” button. It was pretty quietly rolled out, and you can find it under the three-dot menu on any given post in your feed. If there is anything you suspect was done by AI, you can click it and actually report it as “AI slop.” At the same time, LinkedIn also removed its own AI “enhance your post” writing tool and has now replaced that with a proofreading feature instead.
I recently did an online webinar about using AI on social media, as I try to keep up with these sorts of things because it matters. It matters to me and what I’m doing for my clients, but it also matters for small businesses like you and making sure you’re doing the best things for your business. I was a little surprised to hear this particular organization push generative AI images so much, because the people, the public, our potential customers, are growing less and less trustworthy of AI content.
Distrust for AI is Growing
I’m local to Colorado Springs, where right now there’s a lot of discussion going on about a new AI data center they want to build, and there’s been extreme pushback. A lot of that pushback involves environmental issues and the resources it takes to do this. This is also coupled with our utility rates going up, so that’s kind of a whole thing.
But the point is, people don’t want to see everything made out of AI. One of my favorite posts I’ve seen recently was basically a piece of notebook paper that someone had pulled out and written something like, “I have an event coming up this Friday, 5 – 7.” The point was that they would rather go to an event promoted with this handmade flyer than something made with “AI slop.”
People are pushing back. People want to take better care of the environment, and there’s a lot of distrust in AI. When it comes to social media, what I’m seeing a lot is that when people see content made out of AI, they don’t want it. They don’t like it. I’ve been saying this for a long time with reels, people want REAL people. The days of Snapchat filters and perfect hair are gone. The people want the wonderful and imperfect YOU!
Then It Happened to One of My Posts
So that was about two weeks ago. Let’s fast forward to this week. There have been some rumors about MySpace maybe making a comeback, and I am totally here for that. On the post I did about it, I created a carousel, and one of the images was a mockup of what my MySpace used to look like back in the day. I will admit, I did use AI to create. I had actually created this particular image about a year ago, and a lot of that was before I knew as much about the environmental impact of AI. Now, I really limit what I’m using it for when it comes to images, and I haven’t used it to create an image in quite some time.
If you’re using AI for images, you can do that, right? It is a tool. It’s there. It’s something you have available to you. But the point of what I’m saying is that I want you to be aware of the optics. How are people looking at your business when you’re using AI?
Anyway, this was an older image, and really it was just a computer screen that I had done with AI. Everything else in the carousel was pictures I had grabbed of things that my room would have looked like back in the day, with some early 2000s references thrown in. It was about a six-slide carousel. That one particular image was AI, and the others were basically pictures of me that I used as kind of a B-roll backdrop. So, I posted it, and I did not toggle the little AI button that is on Instagram. I probably should have, but honestly didn’t think about it because it was just the one image.
After going back to look at it, I noticed there was now a little disclaimer on my post on both Instagram and Facebook that says “AI Content.” And I was like, oh, well, I guess it was. I don’t really love that because I have been trying to limit my use of AI for stuff, but I understand why they put the AI content label on there. After doing a little more digging around, though, I found there’s a new issue coming up with things being labeled as made with AI.
What Actually Counts as AI?
So many of the tools that Canva has actually do use AI now. The one that I am really frustrated with when it comes to labeling things as AI is Background Remover. Background Remover is pretty much my favorite tool on Canva. It is a technology that predates this big AI boom. It was something you could do in Photoshop and was definitely there before all of this.
Now, there are other tools in Canva that do use generative AI. One that I have used on occasion is Magic Expand. Magic Expand is very obviously AI. It is taking your picture and then adding background or things that aren’t there. I understand labeling that one as AI because AI is actually generating something new. Background Remover is different to me. You are not adding something that wasn’t there. You are removing other pieces of the image. It’s really just very precise cropping.
Think about Snapchat filters. We’ve been using those for years, and they can detect your face, figure out where your eyes, nose and mouth are, and put dog ears on your head or a nose on your face. That kind of technology can use computer vision, algorithms and even machine learning, but the filter itself is following predetermined instructions. It isn’t generating a brand-new face or background for you. That’s different from generative AI taking your photo and creating something that wasn’t there before.
I think that distinction matters because right now we are throwing the word “AI” onto a lot of different technology as if it’s all the same thing, and it’s not.
The Platforms Are Trying to Detect AI, Too
What’s happening right now is that a lot of the social platforms have AI detection or labeling systems, but these tools are not always right. And again, it’s really their own version of AI trying to figure this out, right? So it’s kind of contradictory. If you are of the mindset that you do not want any of your content being AI-generated, now we have this issue where using tools that aren’t actually generating content could still potentially lead to your stuff being labeled as AI.
That is really, really frustrating for me. I hope the platforms find a way around this or give you a way to dispute it if something is labeled as created by AI when it’s not. There are some things that just aren’t the same. Removing the background from a photo is not the same thing as telling AI to create an entirely new image for you.
And yes, people had issues with Photoshop too. Is this accurate? Is it even this real and true? Those conversations aren’t necessarily new, but right now there’s a lot of distrust around AI, and people want to be able to trust what they’re seeing.
Now There Are Reports of AI Watermarks in Text
And this isn’t only happening with images. There are now initial reports that Anthropic, the company behind Claude, is adding an imperceptible, machine-readable signal directly into text generated by models released on or after August 2, 2026. According to the reporting, this isn’t metadata sitting alongside the text. It is actually part of the text itself, so it can travel with the text when it is copied and pasted and may even stick around through some editing. You may have heard about a professor who used this watermark like text on a test (very funny read!).
This is reportedly tied to the EU AI Act and its rules around transparency for AI-generated content. Anthropic isn’t alone here either. Nearly 200 companies had reportedly signed the EU’s Code of Practice on Transparency of AI-Generated Content by the end of July, including Meta, Microsoft and OpenAI. But here’s where I think this gets tricky. Let’s say I write this entire blog post myself, which I did, and then I put it into Claude and ask it to proofread it. I wrote the content. Claude didn’t. But if Claude processes it, could that machine-readable signal now be attached to something I wrote myself?
Based on the initial reporting, potentially. The same issue could come up if you use Claude to translate something, summarize your own writing or reformat it. So even if one of these watermarks is eventually detected, that doesn’t necessarily mean Claude wrote the content from scratch. It could simply mean Claude was used somewhere along the way.
AI Detecting AI on Social Media?
It isn’t a perfect system on the other side either. Heavy editing, paraphrasing, translation, short passages or combining Claude’s output with other writing may make the watermark undetectable. And there’s another pretty big piece to all of this, the detector doesn’t fully exist yet. Anthropic has said it is still working to enable users and other third parties to actually detect the watermark. So this does not mean Instagram or LinkedIn can suddenly scan your caption today and know that you used Claude or ChatGPT. This is more about the infrastructure being put into place and where AI detection may be headed.
There’s already been pushback, too. Some people are concerned about exactly what I mentioned above: human-written text potentially being flagged after Claude simply edits it. There have also been reports of users canceling subscriptions because there isn’t an opt-out. This is all very new, so I am sure we’re going to learn more about how it works. But it’s another piece of this whole AI conversation that businesses need to at least be aware of.
So What Do You Do With All of This?
If you’re using AI, all I would caution you to do is think about it. Think about the impact it’s having on your potential audience. If your potential audience doesn’t care, then that’s what works for you, right? But I feel like I’m not doing a proper service to you if I’m not giving you all the information that I have.
Is some of this going to change? Yes, probably. AI is evolving every single day, and the social platforms are evolving every single day. I don’t love things that are totally AI where everything starts to look exactly the same. You have to have good prompts when it comes to AI anyway, and not having a good prompt makes your content look like everybody else’s. And that’s really just not what we want.
So if you’re using AI, tell me how you’re using it. If you’re super against it, please share that too. Stay tuned and make sure you follow me on Facebook and Instagram, because I’ll continue to share more information about AI and using it in an effective manner for your social media.
*All AI images used in this article were previously made and have been repurposed for the purpose of this article.