Authenticity
in AI

Foreword

The Standard We Are Missing

I am a marketing and branding professional, not a researcher. I want to say that upfront, because it matters for how you read what follows. I put this survey together because we kept having the same conversations in our network — about AI, about what comes next, about what it means for our work, our voice, and the industry we have all built careers in. I wanted to continue those conversations in a way that was actually grounded in something. So the sample is niche. That is the point.

This is primary research from a practitioner. Not peer-reviewed. Not a census. Take it in that spirit.

What I found when the data came back was not what I expected. I expected a story about technology adoption — who is using what, how much, how fast. And that story is in here. But the more interesting story is underneath it. It is a story about the relationship between tools and voice. Between efficiency and integrity. Between doing the work and doing the work well.

AI does not compromise your brand. It does not compromise your voice. It does not compromise your authenticity. But you compromise all of those things when you use a tool without learning how to use it responsibly.

The intersection of analogue and digital — where voice meets tool

We all know what AI slop looks like. We scroll past it every day. The LinkedIn post that is three paragraphs long and says nothing specific. The email that sounds polished but could have been written for anyone. The caption that uses the word "leverage" three times. We feel, in our gut, that someone outsourced their judgment. And it erodes our trust in them — not in the tool, in them.

That erosion is the real subject of this report.

52% of our respondents said AI's effect on brand authenticity "depends on context." That is a sophisticated answer. It is also a convenient one. Depends on context without defining what that context is, is not a position. It is a gap where risk lives. This report is an attempt to close that gap — with data, with practitioner voices, and with a framework that hospitality professionals can actually use.

This is the second in what I hope will become an ongoing annual series. In 2023 we looked at generational gaps in the workforce. In 2026 we turned to the question that has arrived in the middle of that conversation: artificial intelligence. The two topics are not separate. The same generational differences that shape how people communicate, lead, and find meaning at work are now shaping how they adopt, resist, and navigate AI. This report is the next chapter.

Section 01

What People Are Actually Doing With AI

75% of respondents use AI at work. The ones who do save an average of 5.8 hours a week. That is the headline. What gets interesting is everything underneath it.

Text Generation Dominates

When we look at what the 75% are actually doing, text generation is far ahead of everything else. 93% of AI users are using it to write, rewrite, draft, and edit. Images, data analysis, and automation are secondary uses. The tool has been widely adopted for one specific job: words.

Base: 70 AI users. Respondents could select multiple use types.

The Adoption-Integration Gap

Here is the tension: the average person using AI saves 5.8 hours a week, but most of them are only spending 0–10% of their workday with it. People have picked up the tool. They have not rebuilt their workflows around it yet.

That gap between adoption and integration is where the opportunity lives. It is also where the risk lives. People using AI occasionally for text tasks, without governance, without standards, without a clear sense of voice — that is precisely how content loses its specificity and starts to sound like everyone else.

Openness and Culture

71% of AI users say they are always open with colleagues about their use. Only 3% say they never disclose it. That transparency is genuinely encouraging — but the open-text responses add nuance. The standard people apply is not simply "did you use AI?" It is "did you bring judgment to it?" The respondents who said they feel great when a colleague uses AI almost always qualified it: as long as they reviewed it, as long as they added something, as long as it passed the test of actually sounding like them.

Section 03

AI in Hospitality

The human moment AI cannot replace — and should not try to.

The hospitality industry is where this conversation gets most interesting — because hospitality is, by definition, about human connection. The question of what AI should and should not do in this industry is not abstract.

The Comfort Paradox

The same application — chatbots for customer service — is simultaneously the thing respondents are most comfortable with and the thing they react most negatively to. That is not a contradiction. It is a precision problem. The difference between a chatbot that routes a booking inquiry and a chatbot that makes a guest feel like their concern is being handled by a script is enormous. The technology is the same. The judgment about where and how to deploy it is what changes the experience.

Greatest Concerns

When asked what AI applications create the most concern, surveillance and data collection came out on top. 51 of 93 respondents cited AI-driven surveillance or smart room data collection as a concern. Chatbots were second at 46. The pattern is consistent: back-of-house efficiency is welcomed; anything that touches guest data or replaces human interaction is met with resistance.

Where Respondents See the Opportunity

The open-text responses on AI's positive potential in hospitality tell a consistent story: free the people. The most common theme, across all generations and roles, was that AI should handle the administrative and repetitive work so that hospitality professionals can focus on what only humans can do — making guests feel genuinely cared for.

That framing — AI as the thing that protects human connection by clearing away everything that gets in its way — is the most useful frame this data offers. It is also the most demanding one, because it requires intention. It requires knowing what you are protecting and designing the AI integration around that.