Missouri Press Association
Serving Missouri Newspapers Since 1867
I admit I’m a little late to this discussion, but WTH are we going to do about AI? I don’t have any answers, just a lot of questions and a few opinions.
I expect I’ll know a lot more after I attend MPA’s Dec. 10 training on AI. Austin Lewter, a newspaper owner and professor from Texas, will share his expertise. The training is free and available on Zoom for MPA members. Please sign up. We can learn together. Austin promises to help us develop smart and ethical ways to use AI.
On the surface, I want to say a resounding “no” to using AI. It’s a scary beast that threatens our entire civilization. However, practical Peggy knows it can be a handy tool that serves as friendly research and personal assistant. I’m both excited and terrified by the possibilities.
Excited by the possibilities
Truth be told, we already use AI in our everyday work. We use calculators (that’s old school AI for math). We use spell check and grammar checkers. We use digital recorders and transcription apps like Otter. The Zoom app automatically generates closed captions. Our smart phones use predictive text to speed up our messages. We know these services need a human set of eyes to catch the mistakes and to override the wrong.
AI can serve as my personal assistant. At the suggestion of co-worker Katelyn Mary Skaggs, I now ask AI to prepare a packing list when I’m off on a trip. If I input the agenda or itinerary, AI does a good job of suggesting things I pack, like a sweater because meeting rooms are cold or sensible shoes because the agenda includes a visit to a fish hatchery. And don’t forget your laptop power cord.
I can ask AI to find me the best deal on matching pajamas for my grandkids or to generate a Christmas shopping list for my adult child who asks for nothing. When I subscribe to repeat orders for “healthy” snacks, that’s AI at work.
I asked ChatGPT for suggestions to name our newest event, then refined the suggestions by asking for shorter, headline and T-shirt friendly options. AI asks me for clarity. Peggy, do you want this title to be formal or more casual or even hip. (Oh, AI, I always want to be hip.)
We ask AI for better email subject lines and even for headline suggestions. Of course, a real person makes the final call.
Almost every product we use offers an AI component. Adobe Acrobat offers to summarize documents. Photoshop has Firefly, a text-to-image generative feature. Tell Photoshop you want an image of a bird, and it uses AI to make one. Adobe Stock is filled with AI-generated images.
Can I ask AI to use the AP Stylebook to review or edit a news story? Yes. Is it fast? Yes. Will it do a good job? Maybe. Does AI know Eureka, MO, isn’t the same as Eureka, CA? Maybe, if I tell it.
Does it know there are two ways to spell a common family name, and it’s easy for reporters to make a mistake? Probably not.
Terrified of the possibilities
When thinking about AI, I’m reminded of a lesson I tried to teach my own children: Just because we can doesn’t mean we should.
My cousins, who work in the AI industry, said scientists were not given the same advice when tasked to develop the AI science. Can we? Yes. That’s been proven. But now that the AI genie is out of the bottle, it seems a little late to ask, “should we.”
AI is becoming better and faster and harder to detect. It’s pretty clear that the uptick of historical videos on my Facebook Reels is generated by AI. YouTube has lots of AI generated content with images that almost match what the AI voice is describing.
I recently learned that tons of employment interviews use AI. Prospective employees use AI offscreen to smoothly answer questions from interviewers. And some companies are using AI-generated staff to ask the questions. These job interviews are conducted in the digital space – Zoom or Teams, for example, and the responses, face and mannerism all could be AI fakes. It’s possible that no human exists on either side of the interview. It’s so prevalent that Google and LinkedIn offer advice on how to detect an AI interview.
As news organizations, are we conducting news interviews on Zoom or Teams? Are we sure the source on the other side of the interview is a real human? We are pretty sure responses to “I only respond to email”-questions are generated by AI. But what if the whole thing, friendly Zoom banter, witty chat messages and all, is an AI fake? Can we tell?
I guess the real fix is to insist on in-person interviews, and to look for balance between using AI to help and letting AI take control. As my friend Bryan Jones told me: AI, like a hammer, is only a tool. A hammer can be used to build a wonderful home. It also can be used as a murder weapon. I want to protect our journalism and our readers from any sci-fi-like AI harm. We need to be smart in using a new tool.
Stay tuned.
