April 17, 2026
What Meta’s AI Avatar Tells us About the Future of Workplace Trust
Meta is building a photorealistic, AI-powered version of Mark Zuckerberg. Trained on his mannerisms, tone, publicly available statements and his own thinking on company strategy. The goal is apparently that employees “might feel more connected to the founder through interactions with it”.
Ignoring the obviously dystopian nature of this endeavour, let’s ask the deeper, human question: what does this actually reveal about how we understand connection, trust and the human contract at the heart of employee experience? And what does this tell us about where EX design might be headed?
Meta’s AI Avatar Cannot Code Connection into Existence
The premise of the AI Zuckerberg is seductive in its logic. Employees want to feel connected to leaders who are busy and cannot connect with thousands of people in their businesses. Technology can bridge that gap. Easy fix.
Except that it’s not, because the type of authentic human connection that creates psychological safety, discretionary effort and the willingness to stay when times are tough, isn’t built through access to information in a familiar voice. If it were that simple, then company intranets and CEO video messages would have done the trick years ago already.
This kind of human connection cannot be coded by software engineers because, at its core, it’s human, vulnerable, messy, imperfect and unpredictable. Seeing a real human make a decision under pressure or handle a moment badly and acknowledge it or change their mind in real time – those human experiences build trust and empathy through the very vulnerability and imperfection that AI is trying to optimise out of existence.
Right now, younger employees want to spend more time in the office and at least one in five people reports feeling lonely and isolated on remote and asynchronous teams. The problem is especially pronounced among senior leaders who, according to Gallup’s State of the Global Workplace 2026, score the highest on stress, anger, sadness and loneliness – higher than managers or individual contributors.
This context alone doesn’t make Meta’s AI avatar a good idea. The answer to human isolation and dwindling mental wellbeing isn’t another chatbot – it’s actual community and connection. We cannot simulate that level of proximity and a curated performance of someone’s personality. No matter how photorealistic, it is still just a performance.
Vulnerability is Not a Variable
Along with credibility, reliability and reduced self-orientation, vulnerability is a core component of trust. Sure, AI can churn out consistent messaging, but that’s not going to build trust with individuals who are not able to observe consistent behaviour over time and who cannot connect to a shred of actual humanity in a moment.
Research has consistently shown that leaders who demonstrate authentic vulnerability build deeper psychological safety in their teams. AI will always be on message and will always respond with absolute algorithmic certainty and that is precisely what will make it feel hollow.
Truth That Damages Trust
AI hallucination is a documented, ongoing reality. These systems confabulate, and they generate plausible-sounding content that has no basis in fact, delivered with complete confidence.
Let’s imagine that happening in the voice, face, and mannerisms of your CEO.
The AI Zuckerberg tells an employee something about company strategy that the real Zuckerberg never said, never believed and never sanctioned. Who is accountable? How does the organisation walk that back? Most critically: how do employees ever fully trust the next interaction with the avatar, knowing it might be fabricated.
This is not a hypothetical, but rather an inevitable consequence of deploying generative AI in a context with zero margin for error – the simulation of a named, living human being in a position of institutional authority. The reputational and relational damage of a single significant hallucination in this context would undo years of carefully built employer brand and there would be no way to contain it.
The Existential Threat
At Meta, employees are being pushed to use AI tools as much as possible, with ‘skills baseline’ and ‘vibe coding’ exercises prompting fears over future job cuts in addition to the nearly 20% headcount reduction being proposed. People are facing relentless pressure to prove AI competence, all whilst facing significant, fear-inducing job losses and now there’s an AI version of the CEO to share how the human CEO is feeling about it all.
Whether it’s intentional or not, this sends a signal that says ‘we will invest in AI to simulate your connection to leadership at the exact moment we are using AI to make you all replaceable.’ This simulation of care delivered alongside massive job losses lands as a dismissal of the very humans that made the company the massive success that it has been for nearly two decades now.
We are not going to be able to solve the connection crisis with sophisticated technology. Nobody needs more access to a simulation of leadership. Instead, our people need opportunities to connect as humans, outside of work and share human experiences. They need smaller teams, flatter structures and more leaders who show up imperfectly, vulnerably and unmistakably human – and who give them all permission to do exactly the same.
Deborah Hartung is a leadership & culture transformation consultant, executive coach and author with 25 years of experience working with progressive organisations globally. She writes about the future of leadership, employee experience, and what it takes to build workplaces that are both high-performing and deeply human.
