Crisis communication and AI
Can we use AI in crisis communication? Yes, of course — but only very carefully. When you know what you are doing and what you want to say. When you are the one running the communication, you are prepared and you know what needs to happen. In other words, when you are the one holding the steering wheel.
There are three problems with using AI to make communication decisions during a crisis.
The first is that in a crisis you do not have the capacity to prompt AI well and explain the whole context to it. How you communicate with employees, who all your audiences and stakeholders are, how they relate to one another. What was happening in the company before the crisis, what problem has arisen and who already knows what about it.
AI cannot find its way around this context on its own. And prompting it well, with the full context of your business and the current crisis, costs time and mental effort that you simply do not have in a crisis.
An experienced manager carries a large part of this context in their head. They know who makes the decisions in the company, whom they trust, what might leak, how employees will react, what the media will be interested in, what customers are worried about and where the next problem could come from.
That is why they can reach the right decision faster. Not because they can write a statement faster than AI. But because they do not first have to translate the whole context into a prompt in order to get an answer they will have to judge anyway.
The second reason is a little different. AI usually broadens your view, which can be a great thing when you are looking for solutions. It offers more options, suggests more paths. But in a crisis you do not need more paths. You need to make the right decision about which one to take.
You know what is sensitive for your company. You are a human being and you sense the right emotion when communicating in difficult moments. Your experience lets you weigh many factors at once and choose a path.
And third — sensitive internal information does not belong in commonly available AI tools. When you use them, you have to know what happens to the data you enter and on what terms the provider processes it. So you cannot copy confidential internal information, or the personal data of employees or customers, into a prompt without thinking. In a crisis, when the pressure is at its highest, this is exactly what is easiest to forget.
So where can AI actually help you?
Having it edit a statement. Giving it your communication outputs and a list of steps and asking whether you have forgotten anything. Having it show you the weak points in your arguments, or the questions the media might ask.
If you have the time, you can even use AI to rehearse an interview with a journalist — let it ask awkward questions, test your answers and look for the places where a journalist could push you. After the first few days you can give it the list of outputs and work with it to put together a plan for further communication.
But crisis communication has to be done by you — ideally prepared and trained. AI is an excellent assistant, but nothing more. It can be incredibly fast at processing information and producing outputs. But the speed of the output is not the speed of the decision.
In a crisis the decision is still made by a person who knows the company, understands the situation, can weigh the consequences and knows why they decided the way they did.
