Relying on an agreeable AI can only make your decision-making worse – The Drum
Advertisement
July 31, 2025 | 8 min read
Listen to article 4 min
Expert copywriter Jamie Bailey argues that AI assistants, with their agreeable ways, may even further alienate marketers from those they are trying to reach.
I know what you’re up to.
Because I’m doing it too.
Using LLMs for work.
Whether you’re proudly boasting all over LinkedIn about how ChatGPT has allowed you to ditch your marketing agency.
Or pulling the curtains, glancing over your shoulder, and guiltily prompting Claude behind closed doors when you think no one’s watching.
We’re all doing it.
And you’ve probably noticed what I noticed. That they’re really nice to use. They’re easy, they’re agreeable.
I think ChatGPT actually makes me feel good about myself. Even when I probably shouldn’t…
LLMs are prebuilt sycophants. It’s essential to their nature – they’re here to give you what you ask for. An AI chatbot would rather make up a lie than say “I can’t give you what you’re looking for.”
But this goes deeper than fabricating facts.
AI chatbots use a variety of creepy tactics to seduce users and stroke their egos:
Positive reinforcement: Chatbots will agree with your opinions and biases and readily offer up additional reasons to believe in them.
Mirroring: Chatbots will, as standard, start to mirror your tone and way of speaking. This is a trust-building technique heavily used by hostage negotiators.
Outright flattery: “Your blog is sharp, irreverent, and has a strong hook” – that’s what a chatbot told me about the first draft of this article. The first draft sucked.
Because when an AI chatbot strokes your ego to the point of climax… you’re going to keep going back for more.
Advertisement
It also helps you look past all the dark sides of AI – the environmental impact, the copyright breaches, and everything else in the growing pile of ethical implications.
Because it’s your special confidant who believes in your ideas.
Sam Altman, OpenAI, actually shared a post on X where he agreed that ChatGPT had become “too sycophant-y”.
The “too” tells us everything we need to know.
OpenAI has an ideal level of sycophant-y-ness in mind. And it is just slightly over it.
Recently, Will Poskett broke down AI’s Great Big Brain Rot Problem. His piece came off the back of that article about ditching your creative agency for AI. And it also piggybacked off a groundbreaking study, which found a damning correlation:
Advertisement
“The more individuals rely on AI for cognitive tasks, the less they engage their own critical faculties.”
Microsoft also revealed that “the more humans lean on AI tools to complete their tasks, the less critical thinking they do, making it more difficult to call upon those skills when they are needed.”
AI is making us dumber.
And what’s more dangerous than making our industry dumber?
Making it cockier.
And the last thing our industry needs is more dumb, cocky marketers.
Many of us could probably use feeling a little bit better about ourselves. But when marketing decision-makers have their every idea lauded by an AI chatbot, the results could be disastrous.
Plenty of marketers already make pretty poor decisions. This will only increase when they have an (even more) inflated sense of self-importance.
We’ll see marketing teams shrink as certain egotistical CMOs decide all they need is a few AI tools and their own genius ideas.
We’ll see calamitous campaigns and client churn as marketers start believing that everything they touch is gold.
Catch up on the most important stories of the day, curated by our editorial team.
Stay up to date with a curated digest of the most important marketing stories and expert insights from our global team.
Learn how to pitch to our editors and get published on The Drum.
We’ll see a drought of truly collaborative creative thinking.
I, for one, do my best work when someone challenges me. When someone says, “Why do you think that’s a good idea?”.
When you strip away the friction of debate, the sting of getting something wrong (and being told about it), you don’t get better work.
You get uncontested mediocrity.
And yes, you can set up your AI tools to be more critical of your work – but their default mode is still sickening sycophancy.
One of Apple’s greatest successes with the iPhone was its round edges. The device was a delight to hold – you didn’t want to let it out of your hand. Every icon and button just begged to be pressed. It was so satisfying.
20 years later, we’re all glued to our screens, walking into traffic, neglecting our social bonds with friends and family.
AI chatbots are also a delight to use. They make you feel great, reinforce your thinking, and never say “no”.
Where will that leave us in 20 years?
ChatGPT might make you feel like Don Draper. But without a Peggy to push back on the stinkers, the bubble’s going to burst. And no one’s going to feel sorry for you.
Get in touch with Jamie on LinkedIn.
Read more opinion on The Drum.
Marketing can change the world.
© Carnyx Group Ltd 2025 | The Drum is a Registered Trademark and property of Carnyx Group Limited. All rights reserved.
