Is AI going to take over public relations?

Where AI already works

If you strip PR back to its most basic form, the neatly structured press release, the polite announcement, the generic paragraph that could apply to countless companies, AI can already do that in seconds.

Most press releases follow a set formula. The headline, the summary, a couple of quotes, and the company blurb at the end. That is why I think we will see the first big drop-off in small businesses paying PR agencies purely to produce that kind of material.

If all you need is something factual, with no wider strategy behind it, AI is a fast, free option that will get you most of the way there without spending a penny.

Where I would recommend using AI in PR

To be honest, if I were running a small business, I would not even consider hiring a PR firm just to put out a basic statement. It is not because PR firms do not have value, but because if all you are doing is sharing a short update with no reputational risk, AI can now do that instantly.

Take this example. You run a leisure centre. There is a minor pool contamination, nothing dramatic, but you have to close for 24 hours while it is cleaned. Guests are posting about it online, so you want to stop any wild rumours before they start.

In the past, you might have called a PR agency, handed over £1,000, and waited for a carefully worded email. Now, you could open ChatGPT and type:

“We have had to temporarily close our swimming pool due to a minor contamination incident. Write a short, professional public statement that is reassuring and avoids admitting legal liability.”

Thirty seconds later, you have something perfectly serviceable.

I have had very small businesses contact me about situations almost identical to this, and I have told them outright, save your money and use AI. It is such a basic statement that it is very hard to get wrong. Years ago, I would completely understand hiring a PR firm for something like this because you would want it written specifically for you. But now, if it is that simple, you are better off putting the money back into your business.

If something serious had happened, such as a fatality, that is a completely different conversation. But for smaller incidents like this, you could simply put the statement on your website and social media. If the press want to cover it, they will lift it directly. No stress required.

How I can spot an AI statement a mile away

We are in a strange moment where a lot of agencies are leaning heavily on AI. I understand it when it is for traditional PR, like announcing a new hotel in a chain or a standard “company launch” release. I can see why it is tempting to type the brief into a tool and let it do the work…. But eeek, they’re paying you for your brains.

The problem is that AI has such a distinctive tone that I personally find it very easy to spot. If you are going to use it, you need to be smart, rework the structure, give it a more personal voice, and at least try to make it sound like you wrote it. What astonishes me is when a traditional PR firm charges thousands for an AI-generated statement, sometimes paired with a “plan of action” that is clearly lifted directly from ChatGPT, without even bothering to edit it.

Thankfully in crisis PR, I cannot take that shortcut. I am happy AI is nowhere near replacing the human side of what I do, so I am employed for a little bit longer, ha ha.

AI is changing how people work, and not always in a good way

I am concerned, though, when I see how quickly people turn to it instead of thinking for themselves. I will ask someone a question and I can almost see them straight on ChatGPT, waiting for it to hand them an answer.

The problem is, in reputation and crisis work, you need a human brain to think through every angle of a scenario. It is not just about what is technically correct. It is about how a statement will land emotionally, how different audiences will interpret it, and how a single word could be taken out of context and spiral.

More and more people are now turning to AI for therapy, advice, and reassurance, which is its own kind of problem. I do not know if people have noticed, but AI will almost always agree with you. You can feed it the most extreme scenario, end with “but I am in the right, right?” and it will cheerfully reply, “Absolutely, you are completely in the right.” It is amusing until you realise how dangerous that is. It might be great for your ego, but in real life it sets you up for terrible conversations, entrenched arguments, and an inability to see another perspective.

That is proof, if you needed it, that AI still does not understand empathy or how to challenge someone in a human way. It cannot anticipate the hundreds of tiny shortcuts, subtextual cues, and emotional triggers that make the difference between killing a story and fuelling it. Who knows about the future, because it already frightens me how advanced it has become, but for now, at least, this part of PR is still safe in human hands.

The younger crowd seem to heavily rely on it

I have also noticed younger professionals in the industry becoming overly reliant on it. In the social media world, it is already making hiring more difficult. Some junior candidates want to use AI for everything, and while I am not anti-AI, I am worried it is making people stop thinking for themselves.

Even during recruitment, I can see it. In job interviews with people under 25, they often ask me the exact same questions at the end, and their answers are so clearly AI-generated. On Zoom interviews, I can see them reading off their screen. If I hear the word “elevate” one more time, I might lose it.

I now have a habit of asking a completely random question I know will not appear if they type “what to expect in a job interview” into ChatGPT. I do it on purpose to get a real answer out of them. You can see the sheer panic when they realise they do not have a pre-prepared answer. Their confidence vanishes, they start scrambling for words, the structure of their reply collapses, and even their tone of voice changes. It is fascinating to watch because it shows how much they rely on having the answer handed to them by a piece of software.

This dependency shows up in client work too. I once had a past client who liked to post long, thoughtful social updates. I would write them in their voice, always sending the drafts over so they could make adjustments and add anything personal. Instead of tweaking them, they would run them through ChatGPT and send them back to me in a completely different tone. My work would suddenly read like a generic corporate newsletter. I eventually had to gently point out that AI captions were obvious and probably not the right fit for what they were trying to achieve. They were offended and tried to deny it, but the tone was impossible to miss.

Where AI still falls short

The biggest gap is relationships. AI cannot call a journalist and get them to run a story. It does not know which editors will be receptive to your angle, or which ones will use it as an excuse to attack you.

It also cannot access the more human parts of the job. It cannot overhear a comment that changes your entire strategy, or pick up on subtle tone shifts in a journalist’s questions that signal a different agenda.

And despite all the hype around “AI listening tools”, they are still far less effective than a skilled human when it comes to monitoring social media. They will scrape trending hashtags but miss the smaller conversations where reputations are often made or broken. Even AI models that can browse the web in real time are only as good as the data they can see. If something has not hit the open internet yet, like when a backlash is still developing in closed groups, it will not detect it.

Crisis PR is a whole other world

Crisis PR and reputational management, is a very different skillset from traditional PR. I have tested AI by feeding it hypothetical crisis scenarios similar to real ones I have handled. It barely covered five percent of what was actually needed.

Crisis PR is not about filling out a template. It is about reading the room, anticipating where the story will go next, and sometimes taking the completely unexpected route to shut it down. If you are being cancelled, the last thing you need is a safe, standardised statement from AI. That is how you end up looking cold, corporate and guilty, even if you are not.

It also happens pretty regularly that I will be sitting having a coffee and overhear a group of friends at the next table discussing one of my clients. I will sit there in silence, listening like a complete creep, because those conversations often tell me more than anything I could find online. They are real opinions from real people, and they help guide me towards the kind of response that will actually land well.

What you see online is only the tip of the iceberg when it comes to public opinion. The rest is out in the real world, people saying “secretly, I actually think this” to their friends over brunch, away from the safety and performance of social media. Those are the conversations that matter, and they are often very different from what you will see in public comment sections.

AI cannot grab a coffee or sit quietly through brunch waiting for someone to mention a client. It cannot catch the unfiltered moments that reveal what people really think. For now, that is still firmly human territory, and I am glad it is.

Where this is all heading

AI will improve at the basics. We will see models trained entirely on press release archives, media style guides, and social copy. They will be excellent at structured, formulaic PR.

But that is not the same as managing a reputation or steering through a crisis. The work that changes public opinion, avoids disaster, or turns a backlash into an opportunity still needs human judgement, intuition, and relationships that you cannot automate.

For now, AI is a tool. The person holding it is still what makes the difference.