AI Is Like Fire. It Can Burn You or Cook a Good Steak.
If you've been wondering whether AI is good or bad for content marketing, you're asking the wrong question.
AI isn't replacing marketers or expertise, and it certainly isn't replacing trust.
For B2B businesses built on referrals and long-term relationships, AI is simply another tool. If used well, it saves time without sacrificing quality. If used badly, it can damage your reputation faster than almost anything else.
That's why the real question isn't whether you should use AI.
It's whether you're still the one making the decisions.
AI Is Like Fire. It Can Burn You or Cook a Good Steak.
That line came from Dennis Yu, one of the sharpest digital marketers working today, during an interview I did with him on my podcast. He said it almost in passing but it's stuck with me ever since because it's the most accurate description of AI in content creation I've come across.
Most people are either terrified of AI or completely reckless with it. The terrified ones avoid it entirely and wonder why their competitors are producing twice the content in half the time. The reckless ones set it and forget it, pump out content they haven't read, and slowly watch their audience's trust evaporate.
Neither is the right approach.
Why I’ve never hidden that I use AI
I've used AI as part of my content process since it arrived. My clients have always known. That's a deliberate choice and here's why.
I've preached since day one that AI is a tool, not a crutch. My clients trust me because of my honesty and they know that being highly productive with my time isn't optional for me. My family is my number one priority, which means every hour I spend working needs to count. AI helps me do more without sacrificing quality or the human judgement that makes the work worth anything.
There's also a fear of the unknown problem. AI is still relatively new and it's getting a genuinely bad reputation because of how some people are abusing it. I don't want my clients to be afraid to ask me questions about it. I don't want them associating AI with the bros out there who set it, forget it, and think nobody will eventually notice. Because people do notice. Faster than you'd think.
How to spot AI-generated content in three seconds
If you spend any time on LinkedIn you've already developed an instinct for this, even if you haven't named it yet.
The tells are consistent. Em dashes everywhere, AI has an inexplicable love of the em dash despite it being genuinely uncommon in natural written language. Sentences that follow the same rhythm over and over: short declaration, slightly longer declaration, short declaration again. The classic "it's not this. It's that." structure repeated until you want to close the tab.
The biggest tell though is when the content suddenly doesn't sound like the person posting it. If you know someone personally and their LinkedIn posts read like a corporate brochure, you know immediately. The personality has been scrubbed out and replaced with something technically correct but completely hollow.
And then there are the comments. "Drop a comment below if you want me to send you the full guide." That works on Instagram. On LinkedIn it signals immediately that you're dealing with a lead generation machine, not a person having a genuine conversation.
Why AI should be like Jarvis, not Ultron
When clients are nervous about AI being involved in their content, I ask them one question. Have you seen The Avengers films?
Because the relationship between Tony Stark and Jarvis is exactly the right frame for understanding how AI should work. Jarvis does the heavy lifting. The life admin. The things Tony Stark could not possibly manage himself given everything else on his plate. But Jarvis doesn't make the decisions. He doesn't set the strategy. He doesn't replace Tony's judgement, his values, or his relationships.
And importantly, Tony always knows what Jarvis is doing.
That's the distinction. In The Avengers: Age of Ultron you see what happens when an AI operates without proper human oversight. Ultron doesn't ask permission. He doesn't check in. He just executes, and the consequences are catastrophic.
The lesson isn't that AI is dangerous. The lesson is that supervision matters. Especially when you're still learning what the tool can and can't do well.
I tell nervous clients that in the first few months I'm happy for them to approve every piece of content. It's their profile and their reputation after all. If the AI isn't getting it right, we tweak it. That's the learning process. AI is designed to be helpful. It is not designed to be set and forgotten.
The Burn: What happens when AI replaces human judgement
I haven't had this happen personally, but I've seen it play out for others. AI generating an image with garbled text that directly contradicts the brand's stated values. AI responding to comments in exactly the same way every time until it looks like a bot, because it is one. AI saying something genuinely offensive to a follower because it has no idea who that person is or what they've been through.
Every one of those moments does the same thing. It kills the trust the audience had for that person. And trust, once gone, is extraordinarily difficult to rebuild on a platform like LinkedIn where your professional reputation is the entire point.
The Steak: what good AI content looks like
For me it's whiteboard videos. I genuinely dislike graphic design and video design takes me forever when I do it manually. But the AI platforms that specialise in it, given the right prompt, execute beautifully. Canva's AI tools for example will get a design to a point where I only need five minutes of tweaking instead of an hour and a half of building from scratch.
That's AI cooking a good steak. It does the heavy lifting. You apply the judgement, the taste, the finishing touches. The output is better than if you'd done it alone and faster than if you'd done it manually.
The 80/20 rule applies here perfectly. AI produces 80% of the work. You, the senior editor with actual expertise and an actual personality, provide the final 20%. That 20% is the difference between content that sounds like you and content that sounds like everyone else.
Why AI Slop damages trust
If you produce AI slop, you lose trust. Your leads and referrals will dry up. It's that simple.
82% of people can now spot AI generated content. The audience most likely to spot it and least likely to enjoy it are people aged 45 to 65. That's the exact demographic making most B2B purchasing decisions. Publishing hollow AI content to reach that audience is a self-defeating strategy dressed up as efficiency.
So Should You Use AI for Content Marketing?
Yes, absolutely. But only if you're prepared to stay involved.
Think of AI as your first draft, not your final draft. It can organise your ideas, speed up your writing, summarise research and help you overcome a blank page.
What it can't do is replace your judgement.
It doesn't know your customers and hasn’t spent years building your reputation.
And it certainly won't be the one sitting across the table if a client asks, "Did you actually write this?"
The best way to use AI without losing your brand voice
Stop typing your content briefs and start talking them.
Voice memos. Dictation. Whatever you have available. Talk through your answer to the question you want to address, exactly the way you'd explain it to someone over coffee. Then give that transcript to your AI tool of choice.
Here's why it works. When you write something you write differently to how you speak. Your written voice is more formal, more considered, more careful. Your spoken voice is how you actually think and how people actually experience you in person. When you build a business on trust and referrals, your audience is going to meet you eventually. They'll hear you speak. And if your written content sounds nothing like the person they meet, the disconnect is jarring.
Voice memos close that gap. They give AI the raw material of your actual personality to work with. The sentence structures. The specific words you reach for. The tangents and the emphasis and the rhythm of how you actually think.
That's how AI cooks a good steak. You give it the right ingredients and then you stay in the kitchen.
FAQs
-
Yes, provided a human remains responsible for the strategy, editing and final approval. AI should support your expertise, not replace it.
-
Increasingly, yes. Many readers notice repetitive sentence structures, generic language and a lack of personality, particularly in B2B industries where trust matters.
-
Not if you use it well. AI becomes a problem when businesses publish content without reviewing it or adapting it to their own voice.
-
Start with your own ideas. Record voice notes, dictate your thoughts or write rough drafts first, then use AI to help organise and refine them.
Sources
1. 82.1% of Americans can spot AI-generated content at least some of the time, Hookline 2025 AI in Content Marketing Report, https://hookline.io/blog/ai-content-marketing-report-2025
2. 52% disengage when they identify content as AI-generated, Edelman Trust and AI Report 2025, https://www.edelman.com/trust/2025/trust-barometer
3. 95% of content marketers now use AI in their content workflow, Content Marketing Institute 2026 B2B Content Marketing Benchmark, https://contentmarketinginstitute.com/articles/b2b-content-marketing-research/