I’ve just come back from Prague, where I had the privilege of sharing the Reimagining Tomorrow 2026 findings with colleagues from across our profession at the Global Alliance European Summit. I want to be honest with you about something. The longer I sat with this year’s data, the more one uncomfortable thought kept tapping me on the shoulder. We are getting better at AI. But are we getting better at the right things?
Let me start with the good news, because there’s plenty of it. For the first time, we can compare two consecutive years of global data, and the needle is moving. Involvement in responsible AI activities is up. Ethical confidence has strengthened. We’re talking to our stakeholders about AI more openly than at any point before, and more organizations are putting real frameworks in place. That’s genuine momentum, and we earned it. So why am I not entirely comfortable? Because momentum isn’t the same as direction.
What our own priorities are telling us
This year, we added a question I’d wanted in the survey for a long time, an annual tracker that asks the profession one simple thing: what’s your single biggest area of AI focus for the next two years? Globally, the answer was AI literacy and workforce upskilling at 35.5%, then responsible AI governance and ethics at 22.3%, then agentic AI at 12.1%.
Now, I don’t read that as a problem. I read it as honest. It tells me exactly where we are. We’re still learning the tools, building our ethical awareness, finding our feet in an AI-enabled world. That work matters, and I’d never wave it away. But here’s the thing I keep coming back to. Literacy on its own doesn’t create influence. You can be the most fluent person in the room about how a tool works and still have no say in how it gets governed.
"We're becoming more visible in the conversation. But visibility isn't the same as influence. The real question is whether we can convert capability into leadership."
The paradox at the heart of the data
So let me put the uncomfortable number on the table. Involvement in developing responsible AI guidelines has climbed to 57.3%. More than half of us now have a hand in shaping the rules. Wonderful. Yet only 8.6% report leading formal responsible AI structures, and just 10.4% of organizations have handed primary responsibility for responsible AI to the communication function.
Read those two numbers side by side, and you see the paradox. We’re being invited into the room in growing numbers. We’re rarely being asked to lead once we’re there. If we treat professional development as nothing more than tool training, we’ll become extremely good at being competent passengers in a conversation somebody else is driving. That’s not the future I want for us. I don’t believe it’s the future the evidence points to either.
Literacy is the floor, not the ceiling
So how do we close the gap between turning up and leading? My answer is that literacy must be the floor we stand on, not the ceiling we reach for. The investment we’re making in it is sound. It’s also not enough on its own. It needs three companions, and the 2026 data suggests we’re underdeveloping all three.
Governance and disclosure literacy.
Knowing how a tool works is one thing. Understanding how AI should be governed, who should be accountable, and what an organisation owes its stakeholders in terms of disclosure is another matter entirely. With 25.4% of organisations disclosing nothing about their AI use and 20.3% having no formal accountability for AI ethics at all, this isn't an abstract skill. It's the practical fluency that lets you walk into a governance discussion and offer something IT and legal simply can't.
Ethical reasoning that turns into action.
Our ethical confidence has grown beautifully, from 26.2% to 38.8% of us feeling very confident. I celebrate that. But confidence is an internal state. The next step is to convert it into action: to develop, advise on and communicate formal ethical frameworks, rather than relying on gut feeling in the moment. Development must build the muscle of turning instinct into organisational practice.
Plain business acumen.
If we want to lead on the responsible AI dimensions of governance, we must make the case in business terms, at the table, with a clear argument about value and risk. Let's face it, that's a skill our profession has historically under-invested in. The case for our role won't be won on principle alone. It'll be won by people who can connect transparency, trust and disclosure to reputation, risk and the bottom line.
The agentic wave is already here
There’s one signal in this year’s research I can’t stop thinking about. In EMENA, around 34% of respondents named agentic AI and workflow automation as their single biggest focus for the next two years, nearly three times the global average. Globally, agentic AI is already in use or being piloted in 30.9% of organisations.
Let’s be clear about what we’re dealing with, because agentic AI isn’t just a faster version of the tools we know. These are systems that can act, respond, publish and engage on an organisation’s behalf. As those agents step into stakeholder communication, the reputational, ethical and trust implications climb sharply. Yet agentic AI is the single most underserved governance domain in the entire dataset. We’re busy developing for the phase of AI we’ve just lived through, while the next one is already knocking on the door. Festina lente, the old saying goes, make haste slowly. On agentic AI, I’d gently flip it. We need to make haste, carefully, and we need to start now.
Infrastructure on its own won't save us
Here’s something that genuinely surprised me. Australia and New Zealand lead the world on governance infrastructure, with 87.5% framework adoption, the highest anywhere. You’d assume that turns straight into leadership. It doesn’t, not reliably. ANZ also carries one of the highest rates of communication being seen as merely tactical, alongside a stubborn barrier of IT dominance over AI governance. North America tells a close cousin of that story, with frameworks largely in place but the highest rates globally of communication being viewed as tactical, and of having no seat at the governance table at all.
The lesson lands hard for me. You can have the best frameworks in the world and still find yourself on the outside of the decision. What turns a seat at the table into a voice at the table is capability and positioning. The confidence to make the case. The fluency to speak about governance and risk. The standing that only comes from proven expertise. Those are things we develop. They don’t arrive in the post.
Developing against the real threats
There’s a defensive side to this, too and we shouldn’t look away from it. When we asked what threats worry the profession most, the top answers were the erosion of creativity and original thinking at 47.0%, misinformation at 44.3%, and overdependence and declining critical thinking at 38.6%. One in eight, 13.0%, named the devaluation of our own profession, a quiet worry that AI could commoditise what we do if we don’t stand up and defend its value.
I don’t read those as risks to manage. I read them as a development brief. If leaning too hard on AI dulls our critical thinking, then our programs have to sharpen the very things a machine can’t do: editorial judgement, ethical reasoning, reading a room and asking the better question. Let’s face it, the professionals who’ll thrive aren’t the ones who can prompt the fastest. They’re the ones who can do what the model can’t, and who know the difference. Development must protect that, on purpose.
What we need to step up
So where does that leave those of us who design and deliver development, whether through the Centre for Strategic Communication Excellence, our Global Alliance member associations, or inside our own teams? I’d offer four commitments.
- First, we treat literacy as the entry point and deliberately build on top of it, layering governance literacy, ethical practice and strategic acumen so that capability has somewhere to go.
- Second, we develop leadership, not just competence. The profession doesn't need more people who can use the tools; it needs more people equipped and confident enough to lead the responsible AI conversation and own the outcome.
- Third, we are developing agentic AI now, building the oversight, accountability and error-response skills that the next phase will demand of us.
- Fourth, we anchor all of it in what makes us distinctive. Reputation, transparency, stakeholder trust and ethical communication aren't side dishes to AI governance. They're the dimensions that give governance its human and reputational integrity, and they are ours to lead.
Architects of trust
The next couple of years will be pivotal, of that I’m certain. Our long-term value won’t be decided by how quickly we adopt the latest tool. It’ll be decided by whether we step forward as trusted advisors on responsible AI, transparency and organisational trust. If we can position ourselves as the architects of trust in an AI-enabled world, we’ll strengthen far more than our own influence. We’ll strengthen the organisations and the communities we serve.
That, in the end, is a development question. The tools will keep changing; they always do. What lasts is a profession that has built the capability, the confidence and the standing to lead. The Reimagining Tomorrow 2026 report shows how far we’ve come. It shows, just as plainly, how far we’ve still got to travel. So read it. Sit with it. Ask yourself one honest question: what would I need to learn, or help someone else learn, to move from involvement to influence? Then let’s get on with it, together.
