You’d be forgiven for feeling like AI has been an unstoppable force in the last 12 months.
However, the chorus of AI enthusiasts is almost matched in volume by those warning of the pitfalls. Here’s what we think of some of the most common anti-AI sentiments.
“I don’t want AI to replace me”
This is the big one. If AI can produce content, analyse audiences, and work out the right time to send comms, what are marketers for?
The answer is to treat AI as a tool rather than a threat. An AI analysis of your database of leads might highlight 4 under-served segments, but it’s up to you which of them to prioritise. Spotler’s AI models are, in fact, designed to make marketers more valuable; a solo marketer can now handle audience analysis that would previously have taken a full team of data scientists weeks to complete. Our aim is to break the silo effect where data analytics and creative process are separated. When marketers have more direct access to analytics, they can more easily use those insights to fuel their creativity. You’ll be able to quickly ask questions about your data, whether that’s based on a hunch or on a competitor campaign that you really want to steal.
In the headlong rush to add AI to as many areas as possible, we’re finding areas where the human touch still makes a big difference. Take the below example from McKinsey:

“This article is produced with automated text-to-speech, which may not get all pronunciations or voice nuances right.” Really?!
Allowing their website visitors to listen to an article rather than reading it is a good idea! However, the AI-produced result is flat, lacking any of the intonations and emphasis of genuine human speech.
There’s nothing wrong with testing the use of AI in any particular area of your marketing, but don’t assume it will do a better job than you could or even an acceptable job.
One thing that comes up repeatedly in conversations is the fact that AI models appear to have a distinct writing style that’s easy to spot. Much like someone speaking a language that isn’t their native tongue, there are words and phrases you wouldn’t get from a native speaker. So, for now, AI content is quite easy to spot.
We’re also seeing the growth of AI detection tools. Even if the linguistic tells start to disappear, it will still be possible to separate AI-produced content from human-generated.

The best thing you can do is get a sense of how your target audience feels about AI-produced content and adjust AI’s role in your process to meet their expectations.
Quantity over quality
Not got time to write a new blog? Just ask ChatGPT/Gemini/Claude to do it.
A lot of the early excitement around Generative AI was about producing content more quickly and easily. This then led to marketers churning out content in record time, publishing six blogs on a topic where they might have previously taken the time to produce a deeply researched whitepaper. This is driven by the fear that content has become ephemeral, that you will be forgotten if you aren’t sharing something new as frequently as possible.
It’s worth comparing this version of events to how Instagram and TikTok were initially greeted. “Humans now have shorter attention spans than goldfish” was briefly the most frequently uttered sentence on the planet. And yet, for all the chatter around short-form video, the 3-hour-long film Oppenheimer earned 7 Oscars and $968 million. This shows that audience attention depends on the content and the context; people will pay attention to something for as long as it can engage them. Therefore, long-form content can still play a valuable role.
“AI content is too generic to be useful to us”
Large Language Models are essentially aggregators, scooping up huge amounts of existing content and looking for patterns in what has previously earned engagement. This means you won’t get ground-breaking, never-been-done-before content from ChatGPT/Gemini/Claude.
The way to tackle this issue is to have a process that combines human and AI input. Rather than ask your AI of choice to produce a blog, ask it to suggest 5 subheadings. Instead of getting AI to produce a social post from scratch, ask it for 3 new versions of a previous post that did well.
Anything else?
For all the attention AI earned in 2024, we’re still at a very early stage in understanding how it can add value to our marketing efforts. Our recommendation is to experiment with AI as an assistant and be clear about what you are happy for it to do. Burying your head in the sand just isn’t viable; AI will have an impact. Even if you decide that you want to keep all your content and analysis fully in human hands, make that decision rather than getting stuck in indecision.