The AI arms race is no longer just a tech industry problem. It is a multi‑industry race, and it is moving fast. This pace leaves little headspace to consider where the “human‑centred” promise of AI really sits. We are past the point of asking whether to use AI in our work. The real question now is whether we are using it in a way that protects what makes the communications profession human, nuanced and impactful.
A few weeks ago, I attended the London Tech Show, packed with innovation in AI hardware and software and driven by a relentless chase for what is next. What was largely absent was any consideration of human behaviour beyond its replication, or what this wave of AI actually means for end users trying to make sense of it all. Later that day, I attended a talk at Birkbeck, University of London on human‑centred AI – a stark contrast. The discussion highlighted how little space there still is in most tech development for social sciences, psychology, or genuine consideration of how humans think and behave. AI innovation remains predominantly tech‑centric rather than human‑centric, and the gap between those two worlds was hard to ignore.
For those of us in communications and PR, that gap should give us pause.
Keeping it real
Human‑centred AI prioritises human needs, values and behaviour from the outset. It draws on cognitive psychology, behavioural science and sociology to ask questions that pure engineering does not by default: What is the real human need here? Does it support human decision‑making, or quietly replace it?
The key distinction is whether AI automates or augments. Automation replaces human effort. Augmentation enhances it, keeping a human in the loop with better information, interpretation and genuine control. The Stanford Institute for Human-Centered AI describes its mission as advancing AI to “improve the human condition” – a principle that sounds straightforward, but is difficult to embed when speed and commercial gain dominate incentives across industries.
The uncomfortable truth is that much of the AI being rolled out across PR and communications is not being built with communicators or audiences in mind – not at the same time, anyway. It is being built for speed and scale, and sometimes for novelty. That places responsibility on us to decide how it is used to augment our work. Human‑in‑the‑loop is not optional.
Comms is a human profession. Innately.
Effective communication is not built on clever messaging alone. It is built on understanding how people think, what they trust and why they engage. Relationships, instinct and cultural reading are not accidental or easily programmable – they are a craft, ever‑evolving and deeply human.
The data backs this up. A 2025 global study by the University of Melbourne and KPMG, surveying over 48,000 people across 47 countries, found that while two‑thirds of people already use AI regularly, fewer than half are willing to trust it. Crucially, trust has decreased as adoption has increased. Familiarity, it seems, is breeding scepticism, not confidence. For an industry built on trust, that is not a footnote – it is a strategic reality.
Further research published in 2025, spanning thirteen experiments, found that when communicators disclosed their use of AI, they were consistently trusted less by audiences, regardless of how the disclosure was framed. Researchers linked this to reduced perceptions of legitimacy (Research published in 2025). We are not yet in a world where AI‑assisted communications are viewed neutrally. Audiences are developing an AI‑specific “muscle” of perception and judgement, and both are contextual and unpredictable.
The aspiration for AI in PR and comms
The risk is not that AI replaces us. The risk is that nuance is traded for speed, and authenticity for volume, without stopping to consider whether audiences are on board with that exchange.
AI is now commonly used for functional and repetitive tasks: monitoring, reporting, first drafts and research synthesis. Used well, it gives us back time to invest in the relational and strategic parts of communications – the parts that require a real human to show up properly.
Before adopting any tool, the primary question should not be whether it saves time, but whether it makes communications more or less human. Context‑led nuance and genuine audience understanding cannot be outsourced. They are hard‑won experience, and precisely what AI cannot fully replicate.
There is something bigger here, too: The expertise human‐centred AI is meant to learn from – audience insight, trust‑building and cultural literacy – lives in communications teams. We understand how narratives form, how trust breaks, and what makes people tune in or out. That is not a soft skillset; it is what is missing from many AI development conversations. AI can augment this expertise, but only if we stay actively involved.
The opportunity is still open. Communications professionals have a real chance to influence how AI shows up in public discourse – not by rejecting it, but by insisting it augments rather than erodes what makes communication work. By being more intentional about the tools we choose and how we use them, we can help ensure AI strengthens trust, context and connection rather than undermining them. Human‑centred AI will not happen by default; it will happen because communicators choose to lead with judgement, audience understanding and care.
