Testing ChatGPT: Crisis communication without the communicator
How might it help when issues, problems and crises occur?
There is a lot that is being said about the recently launched ChatGPT and particularly how it can provide communication statements and advice. If you haven’t seen it, ChatGPT is an artificial intelligence chatbot developed by Openai. It has been trained to ‘interact in a conversational way’. So what does it mean for communicators?
Many have been suggesting there is an opportunity to use it for some basic communication tasks, which would then allow professionals to focus on more complex problems. Around 18 months ago I asked Kerry Sheehan to share her expert thoughts on artificial intelligence and what it might mean for crisis communication.
So, I wanted to test ChatGPT to see how it may be able to help when issues, problems and crises occur. Would it be able to provide some advice or support, and could it produce sensible statements that could be used with the media and the public?
Many will be aware of the doctor surgery in the UK that sent a message in error just before Christmas. The message told all patients that they had lung cancer when it should have just been a Merry Christmas message. So I asked ChatGPT if it could provide a statement that could be used in those circumstances and the below is what it produced:
![Screenshot of the ChatGPT message which reads: Dear [Recipient] I am writing to apologize for the email that was sent to you in error. It was not meant to be sent to you, and I am deeply sorry for any confusion or inconvenience that it may have caused. I take full responsibility for the error and assure you that steps have been taken to prevent this from happening again in the future. Again, I apologize for any trouble that this may have caused you. If you have any questions or concerns, please do not hesitate to reach out to me Sincerely, [Your Name]](/userfiles/wp/2023/01/chat-gpt_png%20copy.jpg)
After that I asked if it could provide that could be used when a crisis happens and the below was what it provided.
![Screenshot of the ChatGPT question and answer:
AM: Provide a statement for people when a crisis happens [ANSWER]](/userfiles/wp/2023/01/chat-gpt-2_png%20copy.jpg)
I have often said that artificial intelligence can’t yet get the emotional detail that is needed from developing crisis communication. But it is clear that ChatGPT can provide an initial starting point for any issue and crisis communication. You may want to add more specific details to it or to check that the mood and tone are right but these are non-human statements that do show some emotion.
What does this mean for crisis communication? I see it as a positive development. There is an opportunity to use this as a support. It can give some inspiration when your mind may be blank. It can help you turn something around quickly and just check it works for that immediate situation. This can free up the time to develop the more complex longer term communication plan. It definitely means that the focus on copywriting which remains the top task when PR and comms people are surveyed needs to change. It is time to redefine what we do both on a day to day basis and when dealing with a crisis.
Amanda Coleman is a crisis communication expert and consultant, Founder of Amanda Coleman Communication and the author of Crisis Communication Strategies.
This post was originally published on her Amandacomms blog.
Main image by Iurii Motov on iStock.