Phil is a mid-level marketing manager who has become his team’s “content wizard”. Phil’s presentations are flawless, his copy compelling and his ideas seemingly endless. Everyone marvels at his productivity, unaware that Phil hasn’t written an original sentence in months. Instead, he feeds prompts to an artificial intelligence platform, takes the output that “feels right”, makes minor tweaks, and voila.

You’ve met them. Colleagues who present flawlessly articulated ideas they can’t defend. People who present complex topics with confidence until challenged. Executives who respond with vague platitudes when questioned about the deeper strategy behind their choices.

These people are the vibers. And they are powered by AI.

AI invites us to vibe. It seduces us into prioritising AI-generated content that “feels right” over developing our own critical thinking and insights. Vibers can perform understanding without achieving it, using AI to simulate expertise across domains they haven’t mastered. They can fake it but never need to make it.

When AI is used skilfully, it amplifies great ideas and makes us more productive.

AI can be a powerful assistant that can help quickly explore many perspectives of a problem, organise complex information into clearer structures, or help refine narratives. AI can help brainstorm, playfully experiment, and amplify our own thinking.

There can be good vibes: “vibe coding” was coined by AI engineer Andrej Karpathy when talking about weekend projects, describing in a tweet that resonated with many developers.

Here’s an excerpt: “I ‘Accept All’ always, I don’t read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I’d have to really read through it for a while. Sometimes the LLMs [large language models] can’t fix a bug so I just work around it or ask for random changes until it goes away. It’s not too bad for throwaway weekend projects, but still quite amusing. I’m building a project or webapp, but it’s not really coding. I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.”

When insights and direction come from us, AI helps extend our reach and refine our execution. But overreliance is risky, and autopilot is downright dangerous. A subtle yet concerning shift towards “Looks good to me” is emerging.

The student who uses AI to help organise research and generate starting points for their papers now finds themselves submitting AI-generated essays with minimal editing. They begin to mistake the efficiency in completing the task for proficiency. The manager who needs to assess a proposal now uncritically accepts an AI-generated recommendation that “feels right”. They begin to mistake the relief of avoiding difficult cognitive work for intuitive leadership.

Research is already finding that, among high-skilled workers, even with minimal productivity gains, over time they increasingly relied on AI-generated responses. Why think when you can vibe? After all, critical thinking is demanding and requires regular exercise. The ability to evaluate evidence, construct original arguments and detect flawed reasoning is hard won and withers easily. True expertise – understanding interconnected concepts, edge cases, and contextual factors – takes time, attention and care.

It is only going to get harder to resist the seduction of vibing it. AI companies are increasingly focused on getting AI models to “reason” over tasks they are given, as a way to boost their capabilities and potentially broaden their usefulness. Models such as Open AI’s o3 or Anthropic’s new hybrid model Claude 3.7 produce outputs that are compelling but still need to be checked for accuracy and reliability.

The way we work is also likely to contribute to this trend. Workplaces often reward speed over depth; ours is a time of “Move fast and break things”, not “Let’s take a moment to think things over”. We all experience information overload, and staying informed on any subject requires sifting through overwhelming amounts of content. Our collaborations are increasingly remote and rely less on face-to-face or synchronous interactions. In such environments, shortcuts and quick wins are undoubtedly attractive.

Unchecked vibing makes our organisations brittle. We risk not only a subtle homogenisation and a convergence towards average, but more than that we risk mistakes going undetected and unexpected problems arising miles down the track when customers use our products.

We can address the growing vibification of work through upskilling. Specifically, by promoting tech fluency in organisations. The rise in vibing (intentional or otherwise) might be a result of people not understanding how AI models work, how to use them with intention, or the risks inherent in outsourcing their reasoning.

Leaders can play their part by enhancing their AI literacy, so they’re better positioned to create cultures of expertise and productive, responsible tech use. Leaders who speak the language of tech will be able to recraft their organisations with AI in mind.

AI is now a fixture of our working lives – and it has many benefits to offer. By taking the time to understand it while also trusting more in our own thinking, we can experience all the good vibes of this powerful tech without falling foul of the bad ones.

This article was originally published in AFR.