The AI skill no one is talking about

There is so much more you need to know than prompts and tech skills.

This story is brought to you by Ragan\'s Center for AI Strategy. Learn more by visiting ragan.com/center-for-ai-strategyThis story is brought to you by Ragan\'s Center for AI Strategy. Learn more by visiting ragan.com/center-for-ai-strategy
Listen to this article4:27minLearn more

In the last few years, we’ve all been trying to learn how to incorporate AI into our daily work flows. We’ve learned to create tight prompts that bring in context, role and desired outcome. We’ve created custom GPTs that know our preferences and can deliver time after time. Now, many of us are working on building agents that can autonomously do our bidding even while we sleep.

These are important skills that stretch us all as professionals. Learning them is important, let’s not sell them short.

But there’s one skill that is becoming increasingly important as AI continues to inundate us all with more and more thoughts that isn’t getting nearly enough conversation in the AI age.

Editing.

Take this with a grain of salt — editor is in my title. Obviously I think it’s important. But even with that admitted bias, communicators in the time of AI need to be practicing editing skills just as much as they’re practicing their prompts.

What is editing, actually?

Too often when people talk about editing, they’re actually talking about proofreading. The process of looking for typos and grammatic error is important … but I would argue it’s also the part most easily outsourced to AI. AI is pretty good at implementing a style guide. It doesn’t have fingers and so isn’t going to type “pubic relations” instead of “public relations.”

So when we talk about the skill of editing AI, that’s not what we’re talking about.

Instead, what we mean is the process of looking at the content the machine has generated and thinking about it. Deeply. Critically. Analytically.

What questions should I ask about AI-generated content?

The first and most important question to ask about whatever your friendly neighborhood robot gives you is: is this content accurate?

That means manually fact-checking everything it gives you. That may mean visiting the websites it cites to ensure they actually exist and say what the robot purports it does. Or it may mean manually comparing against background documents you’ve given it. In either scenario, you must make sure that what the AI is giving you is factually accurate. If you fail this step, nothing else matters. You’ve lost all credibility.

The second question you should ask yourself is, is this specific?

Specifically what? Well, that’s going to depend on what you’re writing, of course. But AI deals in generalities. It has no lived experiences. It has no friends it can ask questions to. It has no emotions it can infuse into a piece. By definition, it’s giving you the most likely (read: most common, most cliched) answer, simply guessing the next word to string together to achieve the effect you want.

So while AI can give you a good base for a piece, a good editor needs to go in and add specificity to it. That might mean inserting an anecdote. It might mean infusing earnest, emotional language. It might mean something related to your unique company culture. Whatever it is, an editor needs to know what details can be sprinkled into the OK base AI has created in order to make the piece pop.

Just as important is knowing what do I need to cut? AI has a tendency to drone on, filling pieces with puffery and unnecessary words that don’t mean anything. Like a sculptor with a block of stone, a good editor can pare away the waste words. The remaining words will shine through all the more brightly with the dross cut away.

And finally, an increasingly important question: Does this sound like AI? As backlash continues to grow against the use of AI, readers are often scrutinizing any content for gotchas that “prove” it was AI generated. Now, of course, nothing that AI generates is unique. From the em dash to negative structure and beyond, any linguistic structure AI writes was first overused by people. But today, an overuse of some of these tells can cause people to tune out — and for you to lose credibility. So going through and smoothing out any overused AI tics can mean the difference between being believed and being rejected.

What else do you keep in mind when editing generative AI?

Allison Carter is editorial director of PR Daily and Ragan.com. Follow her on LinkedIn.

COMMENT

PR Daily News Feed

Sign up to receive the latest articles from PR Daily directly in your inbox.