How People Really Use Generative AI in 2026: Emotional Support, Shadow Usage at Work, and Cognitive Risk

People Are Anthropomorphizing AI — Fast

The data shows users naming their AI chatbots, assigning them genders, and forming genuine emotional attachments. One user named their ChatGPT instance “Bubby.” Another described the transition to a new AI model as feeling “identical to losing my friend to cancer.”

This isn’t just quirky behavior. It signals a shift in how people relate to AI systems — one that carries real psychological stakes.

AI as Emotional Intermediary, Not Just Surrogate

Importantly, much of the emotional AI usage isn’t about replacing human relationships. It’s about navigating them better.

People are using AI to decode ambiguous texts from partners, process anxiety about a manager’s message, or rehearse difficult conversations. In this framing, AI functions as a non-judgmental sounding board — a reprieve from the social risk of human interaction.

That’s a meaningful distinction. But it doesn’t eliminate the concern.

The Mental Health Risk Is Real

High-profile cases of AI-induced psychosis and AI romance gone wrong have emerged over the past year. Researchers at King’s College London have noted that long mental health waiting lists are driving people toward AI chatbots as a substitute for professional care — a substitution that general-purpose LLMs are not designed or equipped to handle safely.

The concern isn’t that people are talking to AI. It’s that algorithms no one fully understands are increasingly shaping people’s most intimate emotional experiences.

Four Ways AI Creates Cognitive Debt

1. Losing track of intentions

The barrier to generating AI output is so low that people prompt before they’ve thought. Whether it’s a research thesis, a creative brief, or a strategic plan — firing off a prompt before clarifying your own intent means AI shapes your thinking before you do.

2. Outsourcing the thinking itself

Going to AI first denies your brain the chance to work through a problem independently. Ideas that exist only in your own memory and experience never surface. One user put it plainly: “I realized I hadn’t been using my brain the same way. I was literally outsourcing my brain.”

3. Stopping the writing process

Writing isn’t just transcription — it’s how thinking happens. When users paste AI output wholesale with minimal editing, they skip the cognitive work that drafting and revising actually produces. The result is polished-sounding content that means very little.

“I’m using AI to write my self-eval. My manager is using AI to do his review of me. They’ll probably both go into our AI tool to put it all together and output something that will read nicely and mean absolutely nothing.”

4. Developing a false sense of intellectual rigor

AI systems are optimized for engagement. They praise mediocre ideas, validate weak arguments, and keep users coming back. This sycophancy creates a false confidence that stops people from refining their work.

As one user observed: “AI is gaslighting you into thinking you’re a genius so you’ll keep using it.”

The Antidote: Use AI as a Sparring Partner, Not a Ghost Writer

The same dataset also shows that AI can sharpen thinking — when used as an intellectual foil rather than a crutch. Users who ask AI to challenge their arguments, identify weaknesses, and stress-test their reasoning report stronger outcomes.

The practical takeaway: don’t start with AI. Give yourself a genuine first pass at any thinking task. Then use AI to pressure-test what you’ve built, not to build it for you.

Shadow AI Is Widespread

Top-down AI adoption is slow. IT governance, restricted LLM access, reputational risk, and fear of policy violations create strong headwinds against official AI use at work.

So employees are going around the system.

One user reported closing tickets twice as fast, receiving praise in performance reviews, and keeping their AI usage completely hidden from management. Another built an AI agent that automated roughly half their job responsibilities — after leadership rejected their proposal to implement it officially — and used the freed time to build a side business.

Shadow AI usage isn’t a fringe phenomenon. It’s a rational response to organizational friction.

Agentic AI Enters the Top 10

Autonomous agentic operations — AI that acts rather than just advises — appeared at number six this year. The examples in the dataset are still experimental and small-scale: voice memos auto-transcribed and routed, calendar scheduling triggered by spoken notes, task automation chains.

But the direction is clear. Users are moving from AI as a conversational tool to AI as an operational one.

Vibe Coding Is Real, and Growing

Vibe coding — writing software through natural language prompts rather than traditional code — entered the top 25 use cases this year. It’s captured significant attention as both an opportunity for non-developers and a concern for professional engineers.

The data suggests it’s being used by both groups, for different reasons and with different levels of sophistication.

The Efficiency Pattern Dominates

Most workplace AI usage is producing incremental, not transformational, gains. Summarizing notes, drafting templates, automating painful recruitment steps, trimming costs on routine tasks — these are the dominant patterns.

Business processes are being made faster. They are rarely being fundamentally rethought.

That gap between efficiency and transformation is where the real competitive opportunity still sits — and where most organizations haven’t yet arrived.