Whether I’m in a coaching conversation or leading a communication course, I’ve been hit with a new question all year:

“Should I let my team use Al?”

While AI has been around for many years, large language models (LLMs) reached mainstream popularity with the launch of ChatGPT. As more and more companies integrate it into their systems, it’s become unavoidable. Notably, Microsoft (Copilot) and Google (Gemini) now prominently feature their AI’s and encourage their users to embrace them.

As the world around us has started to embrace the technology, we’ve also heard AI horror stories. We’ve seen the research on how Al can hurt our memory, heard stories like the man who poisoned himself using advice from ChatGPT, and all witnessed the lawyer who embarrassed himself using made-up citations go viral.

Despite these drawbacks, we’ve also seen big advances in productivity by people using AI in the workplace. There are undeniable benefits. So, how can we use it responsibly?

I’m going to focus on the use of AI, specifically LLMs, in workplace communication, as this seems to be a popular use case. Let’s unpack the dos and don’ts of using AI at work to communicate.

Dos:

Use AI to check your spelling and grammar.

This is an area where AI tools shine. It can help proof your writing before you send it out to others. It can also recommend words or phrasing that can make your sentences clearer for the reader. If you’re naturally long-winded, use AI to make your emails, documents, or speaking notes concise.

Use it to summarize.

If you have a tendency to write lengthy emails or documents, providing extra information and details, AI can help. Just as we don’t want to read a long email someone else sends us, we shouldn’t be sending long emails with the expectation that the receiver will read it. AI is a great tool to use to summarize your thoughts into a more concise message before you send.

Be careful using it to summarize lengthy documents you receive instead of reading them. AI doesn’t understand what information in a long document is relevant to your role, or how it could impact the decisions you’re making. Take the time to read and understand information critical to your position rather than outsourcing it to AI.

Use AI to help you brainstorm.

AI is a great tool for brainstorming on the fly. Struggling to get started on that report or email? This is where a tool like ChatGPT or Copilot can help get you started. For example, I’ve used it to help me brainstorm titles for my communications workshops. While we rarely choose an AI generated title without changing it a bit, it has been a part of coming up with some of our favorites.

Always re-read anything written by Al before you send it.

AI doesn’t know who your audience is or your relationship with them, but you do! AI is going to write up something that works for a human generally, and that might not work for the person you’re writing to. You will often need to re-write parts of it before you hit send to make it fit the audience and the situation. Check for mistakes too–it’s common for AI to use the wrong word or translate incorrectly to a different language.

Also, take some time to edit it so it sounds more like you. We know that the people we’re communicating with are using AI, but sounding drastically different in writing versus in person isn’t a good look. It’ll start to diminish the value they see in your perspective. If communicating with you feels like talking to an AI, they’ll stop seeking out your expertise and go to another person.

Always double check research and facts.

Surprisingly, Al tools are really bad at this! This is because it doesn’t understand the difference between a factual article and a fictional article. I’ve personally caught it making big errors, and there are lots of stories out there of errors caught by others.

Don’t just trust what the AI says as truth. It’s important to think critically and double check facts and sources. You should make sure that the research it references is from a real academic publication. It’s also worth checking the quality of the research and whether it is still in good standing with the academic community.

Think of it this way—when you send AI gathered information, it’s now linked to your credibility. If you send something wrong, saying “oh AI told me” isn’t a good excuse. They asked you for the information and you are accountable for the quality of the information you send.

Don’ts

Don’t use it to create graphics that impact your brand.

Hire a graphic designer—they are worth it. While it might seem like a great cost saving idea to use AI for all your design needs, it won’t help you stand out. Al will push out the same logo, infographic, or image to everyone. So, if you are creating an ad or updating your website, it’s not worth having the same image as another company.

Never use AI to write a critical email or document.

If you need to have a tough conversation with a coworker, don’t have Al write the email or Teams message. It’s too risky. Workplace communications need to build connection and trust, so write it yourself. An AI written message that doesn’t hit the right tone can quickly damage a relationship with an important coworker.  

Conclusion

Communication is where we build credibility and connection, and only you know your audience. Writing a boss who you’ve worked for 20 years is much different than writing a new boss who just started. It’s important that you understand the value of your personal experiences with another person. All your communications should be informed by your relationships. There is no one size fits all approach to other people.

So, the morale of the story is—yes, Al is fine use in the workplace to communicate. But double check it and read carefully before you send.