In a new essay, Bill Gates makes the case for what he calls “Human Reserved” jobs: vocations we may deliberately choose to keep human-led, even where machines could technically perform elements of the work.

The examples are unsurprising: healthcare, education and the roles that carry us through difficult moments, where empathy, responsibility and human connection matter most.

It is an interesting framing because it challenges the usual AI debate. The question is not simply what machines can do. It is what we should want them to do, and where we still need a human being to exercise judgement, take responsibility and be accountable for the outcome.

So, should communications be Human Reserved too?

My answer is yes, at least in the ways that count.

AI can take the production work

This is not an argument for communicators to reject AI, nor to pretend that the technology will not reshape how we work.

It should.

AI can take care of parts of communications’ production layer: early research, monitoring, transcription, synthesis, first drafts, versioning, reporting and the more repetitive cycles of content development. It can help teams find patterns in large volumes of information, test an argument from different angles, turn a long report into an accessible summary and free people from the routine work that swallows disproportionate amounts of time.

That is not a threat to good communications. It is an opportunity.

For too long, parts of the industry have allowed themselves to be defined by the things easiest to count: the press release, the media list, the activity report, the deck, the volume of output. Those things have always had a place, but they were never the whole job. If we define our value as an output factory, we should not be surprised when the factory becomes easier to automate.

The more useful question is what we do with the time and capacity that AI creates.

The answer should be to move upstream: into interpretation, counsel, risk sensing, stakeholder understanding, narrative design and the hard work of helping organisations decide what is true, relevant and worth saying.

The work that remains human

At its best, communications is about judgement.

It is understanding what matters, who needs to hear it, what should be said now, what should not be said at all, and how a message will land in a complex and often sceptical human environment.

That last point is especially important. A machine can generate a plausible holding statement. It cannot take responsibility for the consequences of issuing it. It cannot sit with an executive team as a situation changes by the hour, assess the political, cultural and commercial implications, or tell a client that the instinct to say more is likely to make matters worse.

Nor can it read a room in the fullest sense.

You can give AI a detailed account of community opposition to a major project. You cannot send it to the town hall meeting, ask it to gauge the mood in the corridor, distinguish performative anger from genuine loss of trust, or understand the local history that makes a technically sound message feel tone-deaf. That work relies on cultural fluency, empathy, context and the accumulated pattern recognition that comes from working with people under real pressure.

The same is true of leadership communications. A leader’s voice should not be a polished composite of what an algorithm thinks someone in their position might say. It needs to be rooted in what that individual knows, believes and has experienced—and it needs to withstand questions from employees, customers, regulators, journalists, investors and communities.

That is communications as a human discipline: not simply making words sound better, but helping people make better decisions about what those words mean and what they commit the organisation to.

Authority in a synthetic world

We are now awash with synthesised content. It is increasingly easy to produce something that looks professional, sounds informed and follows the conventions of corporate language.

That does not mean it is credible.

In fact, the opposite may increasingly be true. As the volume of generic, over-polished content rises, people are becoming more sensitive to the difference between a manufactured opinion and a perspective with real substance behind it. They want evidence, expertise, and a point of view that can be tested.

This is where personal authority becomes more valuable.

Personal authority is not executive self-promotion or an endless feed of leadership posts. It is a distinct, defensible perspective grounded in the problems someone has solved, the body of work they have built, the decisions they have made, and the professional reputation they have earned over time.

Trust compounds when it attaches to a person who can explain their thinking, defend their view in a room, and be held accountable for it. It is harder to manufacture and easier for audiences to recognise.

The role of communications is to help leaders identify the intersection of their genuine expertise, the questions their audiences are asking, and the issues that are genuinely relevant to the business. It is then to help them express that perspective clearly, consistently, and credibly—not to invent authority where none exists.

That distinction will matter even more as AI changes how people discover and evaluate organisations. When AI systems increasingly mediate early research, vendor selection and reputation, businesses need clear, coherent, and verifiable signals across earned, owned, and leadership channels. Trust is not simply a reputation score; it affects whether an organisation makes the shortlist, secures the meeting, or gets the benefit of the doubt.

A more valuable mandate

This is why communications should be more ambitious about its role in the AI era.

Our task is to act as a conduit between what machines can process and what people can understand, believe, question and act upon. We help turn operational reality, governance, safety, product performance, customer outcomes, policy, risk and business strateg, into credible stories that can survive contact with sceptical audiences.

That requires more than content skills. It requires curiosity, commercial fluency, cultural intelligence, scepticism and the confidence to challenge a neat narrative when it is not actually true.

AI is prolific. It is not wise.

Communications professionals should embrace the technology where it improves speed, quality and access to information. But we should be clearer about what it cannot replace: contextual understanding, cultural fluency, empathy, responsibility, the ability to read a room, and reputational judgement earned over time.



Those are not peripheral parts of the job. They are the load-bearing parts.

But I freely admit I am biased.

PR is after all not ER.