It’s safe to say journalists and AI chatbots have not been fast friends. The quick influx of generative AI tools has threatened to replace the work of experienced writers and editors with stale, inaccurate content, called into question the future of media, and upended how people find their information. At the same time, a growing list of media companies have signed content licensing deals with OpenAI. Their writers’ articles will now surface in ChatGPT’s answers and be used to train its underlying model. When The Atlantic signed on, one writer called it “a devil’s bargain.”Can journalists and AI coexist peacefully, maybe even harmoniously? Where can AI be helpful, and what needs to be improved to make it more valuable? To get some answers, we asked our editorial team. Since AI is such a divisive topic, we used an anonymous survey to encourage candid responses. Unsurprisingly, the comments reflect the mixed attitudes many people have about AI.”The effort to combat misinformation disseminated using generative AI tools will be a massive undertaking for the foreseeable future,” says one of our reporters, “It’s pretty concerning.” Some have not yet tried chatbots, for no reason other than they haven’t been compelled. But for those who have, here are a few ways they’ve found them helpful as input to the larger creative process.Editors’ Note: PCMag does not generate articles or images with AI, per our Editorial Mission Statement. 1. Ditching the Drudgework To Focus on Expert Product ReviewsPCMag publishes more than 1,600 technology reviews each year, and each one of them requires many information-gathering tasks before it can even be written. Some of our reviews team staffers use AI to cut out some of the repetitive work so they can spend more time testing products and offering expert buying advice.One of our reviewers uses AI to help sift through pages of marketing materials and find the most relevant information. “I’ll make a combined document of all the marketing materials and spec sheets for a product, [and then] drop it into a tool like Anthropic’s Claude,” he says. “Then, I can query the results faster than hunting for individual specs and technical details. And it makes it easy to double-check everything before it goes into the review.”Entering all those reviews into our systems also takes some bureaucratic work, which AI helps speed up. “I’ve used AI to create several custom tools,” he says. “For example, I have a tool for grabbing photo URLs from PCMag reviews and adding them to product roundups. It’s crude, but makes that sort of time-consuming tasks a lot faster.”2. Quickly Answering Questions That Google Can’tThough not a replacement for a typical Google search, AI chatbots offer another angle from which to come at a subject. “I use Copilot whenever I can’t find an answer I’m looking for in a few web searches,” says one of our staffers. “Or when I’m looking for a tiny tidbit of information and a regular search doesn’t get me.”But using AI for research can be hit or miss. “Sometimes they won’t answer the specific question but instead give you basic info on the relevant topic,” one writer says. Even though the information may be related, that doesn’t mean it’s what you’re looking for.Our writers also find it hard to trust the information, which limits its usefulness. “I feel like I can never take what chatbots are saying verbatim without fact checking,” one reporter says. “I always have to verify the original source.”3. Brainstorming & Finding New AnglesThe best publications provide new information—not just regurgitate what’s already on the web, which is what chatbots specialize in. That being said, they can suggest angles for articles and speed up the initial brainstorming process.”I use AI to brainstorm new articles and quickly understand some of the key nuances of the topic,” says one of our reporters. “It’s never enough to write a full article on or to get a fully baked concept from, but it sparks ideas that I can run with on my own.”ChatGPT and others are also helpful when brainstorming journalism-adjacent tasks, such as a name for a podcast, book, or social media post. 4. Identifying Interview SubjectsOne of our writers uses AI primarily to identify interview subjects, saving her time spent combing through “a heap of academic journals and research papers to find someone who seems to have the right expertise.””One of my favorite uses for AI is helping me find good sources to speak with for reported feature stories,” she says. “It’s honestly one of the hardest parts of reporting a story, but AI can do that search for you in seconds and provide a list of potential sources, which is a massive time saver.”5. Finding Synonyms, New Turns of PhraseAI chatbots can act as a type of advanced thesaurus and help rephrase sentences, making our writing more dynamic.”As an editor, I’m constantly looking for better, clearer, or more illustrative ways of saying things, so prior to AI, I spent a lot of time looking up synonyms on Google,” says one of our editors. “Now, I do most of that with AI because in addition to looking up synonyms for single words, I can ask for “synonyms” of entire phrases or ask for alternative ways of saying something.”One reporter finds AI helpful for playing with language, such as making alliterations.”If I’m trying to make an alliteration with the letter “g”, but one of the words I want to use starts with a different letter, then I’ll ask it to give me a synonym that word that starts with “g,” which I’ve found isn’t as easy with Google,” she says.
Recommended by Our Editors
6. Helping Craft the Perfect HeadlineWriting a headline that’s interesting and accurate, without being clickbait can be tough. Some PCMag editors are using AI as a sounding board in the process of developing them.”I usually prompt AI to give me five variants on a head in different tones (funny, authoritative, etc),” says one of our editors. “It’s invaluable for ideating and riffing on headlines, in particular,” another adds. “You can give it your own headline ideas, and ask it to do a sentiment analysis and figure out what it might imply to future readers,” says one reporter. “Then you can make sure it’s a good fit for the article’s content.”Not all editors find it helpful with this process, however. One says they “haven’t found them to be great at generating headline alternatives,” and another calls the output “utter garbage every time,” but notes that “seeing a bunch of suggestions, even if they’re bad, is a good way to get inspiration and kick off the process of riffing and refining a headline.”To the AI Chatbots: Our Humble RequestsWhile those of us who use generative AI have come to appreciate it for specific purposes, several major limitations hold it back from being the “game-changing technology” the tech industry imagines, at least for written journalism.One of the biggest issues, cited by many of our staffers, is the lack of source citations. Not only does this make it hard to trust the content, but it also hits home: We see our own work quoted, sometimes nearly verbatim, without noting the author’s name or linking to the piece that took us hours, weeks, or even months to write.”Citation, citation, citation,” one writer says. “They’re getting better, but there’s still a lot of room for improvement.””They don’t always link to sources, which I find unethical and unhelpful,” another adds.Inaccurate information, and not getting the facts right “100% of the time” is another major concern (perhaps unsurprisingly, as that is one of the core pillars of journalism.) Even Google CEO Sundar Pichai even admits that “LLMs aren’t necessarily the best approach to get at factuality,” despite adding AI Overviews to search result pages.
Get Our Best Stories!
Sign up for What’s New Now to get our top stories delivered to your inbox every morning.
This newsletter may contain advertising, deals, or affiliate links. Subscribing to a newsletter indicates your consent to our Terms of Use and Privacy Policy. You may unsubscribe from the newsletters at any time.
