In March, the Chicago Sun-Times laid off 20% of its newsroom. Just two months later, it published a “summer reading list” that recommended completely fabricated books attributed to real authors, such as Taylor Jenkins Reid.
Was it satire? Nope. Just an unchecked AI hallucination, pushed live without a single human asking, “Wait… should we?” Or, I don’t know, checking to see if the books exist.
This is the crossroads at which PR now finds itself.
Artifical intelligence has the power to help us move faster, scale smarter, and free up time for strategy. But under reduced budgets, financial pressure, and/or tight deadlines, it’s tempting to hand over the wheel entirely—to let the bot drive the content, the pitch, the post, and the news release.
But that’s where things go sideways.
We’ve entered a new era of automation overload: mass-produced PR pitches, templated thought leadership (has anyone been on LinkedIn lately…oy vey!), and generic content that erodes the very trust and credibility our industry is built on.
So where’s the line between helpful automation and dangerous delegation?
The Double-Edged Sword of AI in PR
On one side, AI promises everything we’ve ever wanted: efficiency, scale, instant drafts, endless headlines, even pitch-ready media lists. It’s like having a tireless junior associate who never sleeps and never takes time off.
(Except for those times when it times out and makes you wait a couple of hours before you can prompt it again. I always yell, “If I wanted to give breaks, I’d hire humans!” Alas).
But on the other side? That same tireless associate doesn’t understand nuance, tone, timing, or relationships—and if left unsupervised, they’re going to start sending mass pitches about your B2B SaaS platform to the beauty editor at InStyle.
Unfortunately, this is exactly what’s happening across the industry.
Artificial intelligence is helping communicators crank out content at breakneck speed, but much of it is… well, garbage. Thoughtless, templated, off-brand, or just flat-out wrong.
The Wall Street Journal recently noted that PR pros are flooding inboxes with generic, AI-generated outreach, leaving journalists frustrated, skeptical, and increasingly unreachable.
We’re not helping our reputations. We’re hurting them.
The problem isn’t the tool—it’s how we’re using it. Or more accurately, how we’re misusing it. Every time someone says “AI doesn’t work,” I respond with, “User error.”
Which is probably not a very nice response on my part, but you do have to know how to use it properly for it to work properly.
So, how do we keep AI in its lane?
By understanding what it’s actually good at—and where it absolutely, under no circumstances, should be left alone.
Let’s start by mapping that out.
What You Can (and Should) Delegate to AI
Let’s give artificial intelligence a little credit. When used correctly, it’s incredibly helpful—especially for the parts of our jobs that are repetitive, time-consuming, or (let’s be honest) a little soul-sucking.
Think of it as the world’s most eager assistant. It’ll never say no. But it will absolutely hand in a first draft at 4 p.m. on a Friday and confidently pretend it didn’t just make up three sources and quote a nonexistent Gartner report.
So here’s what you can let AI help with—safely, smartly, and with the right oversight:
- Content drafting. Use it to draft the initial versions of news releases, blog posts, email sequences, and even podcast scripts—emphasis on the first pass. You’re still the final word.
- Transcription and summarization. Interviews, webinars, meetings—let the bots do the typing. You can skim instead of sifting through the whole thing.
- Headline and subject line variations. Need 15 ways to say “read this”? AI has you covered. Just don’t let it name your campaign “Unlock Synergy for Future Innovation.” It looooves the corporate jargon.
- Background research. Use it to gather company bios, product descriptions, or industry context for a briefing doc. You still need to verify it, but it’ll save you hours.
- Media monitoring and sentiment tracking. AI tools can scan for mentions, surface trends, and even attempt to detect tone. Human interpretation is still required, but it’ll do the heavy lifting.
- Data analysis, trend surfacing, and a little light fortune-telling. Need to understand how a campaign performed across channels? AI can help spot patterns and predict the future—but again, you have to connect the dots.
- Translation assistant. For global campaigns, AI can provide initial translations or help localize messaging, but please, please, please have a native speaker review it before it goes live. Always.
The sweet spot of artificial intelligence is efficiency, not execution. It makes things faster—but not smarter. That’s still your job.
What You Should Never Delegate to AI
There’s a line. And it’s bright. Neon bright. Possibly outlined in glitter.
Because, while artificial intelligence is great at grunt work, it is terrible at judgment. And in PR, judgment is everything.
Here’s what you should never, under any circumstances, let AI run without a human in charge:
- Crisis communications. Timing. Tone. Empathy. Context. All critical. All human. The last thing you want is a bot drafting your apology post—or worse, your holding statement—when everything’s on fire.
- Executive thought leadership. Did I mention how bad some LinkedIn posts are? Your CEO doesn’t sound like ChatGPT. Nor should they. These pieces require authenticity, perspective, and nuance—not generic LinkedIn wisdom and metaphors about climbing mountains. Unless they actually climb mountains.
- Reporter pitches and media relationships. I cannot emphasize this enough: relationships are the currency of PR. If a journalist receives a pitch clearly written by a bot, your credibility is toast. And, speaking from loads of experience, we can tell when your pitch is written by a bot. There are lots of tools to help you with pitch drafts (PodPitch is one of them), but for the love of Thor. Please do a little research and edit them before hitting send.
- Legal or regulatory communications. Do you like fines? Lawsuits? Consent decrees? Then, by all means, let an AI hallucinate your compliance language. For everyone else: get legal eyes on it. Always.
- Final content approvals. AI can draft. It can suggest. It can organize. But it does not hit publish. That’s your job. That’s always your job. Even when a junior-level associate drafts something, it’s your job to oversee the work. AI is a junior-level associate.
- Humor, empathy, and cultural sensitivity. Unless you want your campaign to accidentally offend an entire demographic—or make a joke that only makes sense in 2008 meme culture—leave the delicate stuff to humans.
In short?
If it touches your brand voice, your reputation, or your relationships, it needs a human brain behind it.
The Hidden Price of AI Shortcuts
When we skip that human layer—when we treat artificial intelligence like a strategy instead of a tool—we don’t just risk mistakes. We risk losing trust, credibility, and access to opportunities.
And the consequences aren’t theoretical. They’re happening right now.
According to the aforementioned Wall Street Journal article, journalists are being flooded with formulaic, AI-generated emails. They’re frustrated. They’re ignoring inboxes. And they’re starting to block the worst offenders entirely.
Journalists and bloggers and podcast hosts and influencers can spot artificial intelligence from a mile away. Generic greetings. Over-polished language. Irrelevant pitches. It doesn’t take much to recognize that a bot wrote it, and once they do, your name goes straight to the “nah” pile.
AI mistakes can and do go public. Whether it’s a hallucinated quote, a made-up stat, an inappropriate tone, or an entirely fake book list, AI errors get shared. Screenshotted. Mocked. If your brand ends up in one of those threads? That’s not the kind of reach you want. Never become the story.
And, when everyone is using the same tools and copying the same prompts from the hustle bro on LinkedIn, you become…forgettable. Without oversight, the output becomes eerily similar. And sameness is the enemy of storytelling.
So yes, artificial intelligence can make your job easier. However, it can also make your brand appear careless, your team appear lazy, and erode your credibility overnight.
How to Build in Human Oversight
Let’s get something straight: using AI responsibly doesn’t mean slowing everything down or adding 17 new steps to your process. It just means not skipping the step where a human double-checks that the AI didn’t accidentally recommend “War and Peace” for preschoolers.
Oversight isn’t bureaucracy. It’s brand protection.
Here’s how to build it in without breaking your momentum:
- Train your AI like you would an intern. Remember, AI is your junior associate. You wouldn’t let them publish something without your oversight. You wouldn’t let them send pitches without your review. Same goes for your AI. Provide examples, brand voice docs, and prompt libraries. The better your inputs, the smarter your outputs.
- Always, always review before publishing or sending. This is non-negotiable. Every AI-generated draft should get at least a once-over from a human who knows what “off” looks like. It’s not that AI tries to mess up—it’s just very confident when it does. Even when you question it, it will double down. So check everything.
- Create a clear “AI can and cannot” policy. Document what tasks AI is allowed to help with—and which ones are strictly human territory. This helps your team move fast without overstepping.
- Lastly, build feedback loops. If AI makes a mistake (and it will), feed that back into your process. Update your prompts, add examples to your training library, and keep refining.
The Red Flags Checklist
Before anything goes out the door, run through this quick set of questions:
- Did the AI fabricate a stat, source, or quote?
- Does it sound like a robot in a blazer?
- Could it offend or alienate someone? (
- Is the tone aligned with your brand and your audience?
- Would you open or respond to this if it hit your inbox?
- Has it been run through an AI and plagiarism checker? Grammarly has a nice add-on that allows you to check all your work first.
Trust your instincts. If it feels off, it probably is.
What Happens When AI Works
We’ve spent a lot of time discussing what not to do, so let’s shift gears and explore what it looks like when artificial intelligence is used effectively.
It is possible. It doesn’t involve trying to pass off a ChatGPT draft as a final deliverable. (Please. Don’t be that person.)
Here are a few real-world examples of what smart, human-guided AI use can look like in practice:
Assisted Pitching That Still Sounds Human
An agency used artificial intelligence to generate base pitches for a product launch and then had the team personalize each one with a relevant hook tied to the reporter’s beat. They researched the articles the AI suggested as the prompt, read them, developed their point of view, and then massaged each pitch to be highly personalized.
The hard work was already done by the AI—the person to pitch, their contact information, when they typically responded to emails, recent stories they had written on the topic, and how relevant the pitch was to them. So, all the agency had to do was make sure the pitch was personalized.
Want to know the result? Their pitch prep time decreased by 40%, and response rates reportedly increased by 2.5 times compared to previous campaigns. Boom!
Localized Messaging with a Global Assist
A nonprofit organization used artificial intelligence to analyze thousands of social media comments from a global health awareness campaign. The AI grouped common themes by region, highlighting, for example, that Latin American audiences were concerned about access, while European audiences focused on stigma.
The heavy lift—sifting through overwhelming volumes of user-generated content and organizing patterns by geography—was handled by the AI. Human strategists then stepped in to develop messaging that was culturally appropriate and emotionally resonant for each audience segment.
The result was a 22% lift in engagement across key regions compared to the previous year’s campaign. The AI spotted what mattered. The humans made sure it landed.
Research at Scale, Without the Slush
A communications team supporting a thought leadership campaign for a B2B executive used artificial intelligence to compile a comprehensive briefing. It included competitor positioning, recent executive quotes, top-performing content formats, and whitespace opportunities in the market.
The strategist reviewed the output—double-checking sources, validating insight quality, and highlighting what aligned with the executive’s strengths. With the artificial intelligence handling the foundational research and organization, the humans focused on storytelling, voice, and messaging hierarchy.
Their result was that a five-hour research task was completed in under one hour, freeing up time for creative development and stakeholder alignment. No hallucinations. No fluff. Just fast, focused insight-gathering that let the human strategist shine.
When you use artificial intelligence to scale inputs, but keep human ownership of outputs, magic happens. You get speed and substance. Insight and nuance. Consistency and creativity.
AI Is a Tool—You Are the Strategy
Artificial intelligence doesn’t know your audience. It doesn’t understand your CEO’s peculiar fixation on nautical metaphors. It can’t sense when a pitch feels off or a headline lacks punch.
That’s your job.
Artificial intelligence can make you faster—but only if you stay in the driver’s seat. It can help you scale—but only if you keep a human hand on the tone, the truth, and the trust.
Because in PR, relationships are everything. Make sure your AI helps build them—not burn them.
To Learn More
Make sure you follow us on LinkedIn, TikTok, and Instagram. We’re posting tips, how-tos, cheat sheets, and more every week.
Also join us in the Spin Sucks Community. It’s a great place to brainstorm, get ideas, and collaborate with your peers.
© 2025 Spin Sucks. All rights reserved. The PESO Model is a registered trademark of Spin Sucks.
