 |
| Ronn Torossian |
For decades, the discipline that kept public information honest wasn't a policy—it was a person. A reporter would call to verify a claim before running it, and that single check acted as a correction cycle: stale facts got caught, updated releases got flagged, and outdated pages quietly stopped being the thing people relied on.
AI search tools have removed that check. ChatGPT, Gemini, Perplexity and Google's AI Overviews pull together answers from whatever content they consider trustworthy, and they present that answer with no visible source, no byline and no way of knowing whether the page behind it is still current. The person asking the question sees a confident answer. They don't see a dateline, and they don't get a callback. That shift changes what "keeping information accurate" means for Public Relations—the job now includes making sure the right version of a fact is the one these tools can find, not just writing the fact correctly the first time.
The problem is drift, not error
Most conversations about AI search today focus on format: clear headers, simple definitions, easy-to-scan comparisons. That work matters, but it assumes the underlying content is still true. It often isn't. Product specs change. Feature counts get revised. Requirements get dropped. In one recent client engagement, a device-compatibility page had been replaced by an updated version months earlier—different device count, different requirements—yet the older page was still live and still fully able to be pulled into an AI-generated answer as though it were current.
The content wasn't wrong when it was published. It became wrong later, quietly, the way most product information does. The difference now is that nothing forces a re-check. A traditional search engine points you to a page and lets you judge it. An AI tool skips that step and just tells you the answer—with the same confidence whether it's current or six months out of date.
Why this sits with PR, not IT or SEO
Three reasons this becomes a PR responsibility rather than a technical one:
PR owns the narrative timeline. Comms teams are usually the only ones who see the full arc of what's been announced, updated and replaced for a given product—even when the pages themselves are maintained elsewhere.
Fact-checking has always been PR's craft. Verifying a claim before it goes out, and correcting it after, is the same instinct that's protected client credibility with reporters for years. AI search just raises the cost of skipping that step.
The fix is usually a judgment call, not a technical fix. Deciding whether to retire an old page, how to redirect it, and what to say about why it changed is a communications decision, not an IT ticket.
Verify — and then verify again
Truth has always mattered in PR. A few habits worth carrying into this new environment:
Check the current source before republishing anything. Don't rely on the original brief, the last version that performed well, or memory—check what's actually true right now.
When a fact changes, retire the old version instead of letting it compete. The way we used to take down an old press release, we now need to make sure an old page doesn't stay live alongside the new one.
Write down what changed and why, and keep that note with the content. A short line—what was updated, when, and from where—gives the next person editing it the context to catch the next change before it becomes a problem.
The stakes are higher than a testy follow-up email
The old cost of outdated information was bounded: a correction, an awkward exchange with a reporter, maybe a small embarrassment. The new cost is bigger. A stale page can end up baked into an AI-generated answer indefinitely, with no expiration date and no warning to the person reading it. It's not just a risk of one bad headline—it risks becoming the permanent, invisible answer to a question a customer, partner or caregiver asks about a client's product.
That risk isn't limited to sensitive categories like healthcare or finance, either. Any product with a version history—pricing, features, compatibility—carries the same risk the moment two versions of the truth are both out there for these tools to find.
Raising PR's bar
None of this is an argument against doing this work—it's an argument for who should be doing it. Technical teams can format a page so it's easy to read. Only PR has the standing habit of asking whether a claim is still true, still sourced, and still worth standing behind—and of taking responsibility for fixing it when it isn't.
Agencies that build fact-checking into the front of this process, not as a fix applied after something goes wrong, will be the ones whose clients get represented accurately when someone asks an AI tool a question about them. The alternative is finding out the way most of these problems get found: by accident, hoping it's caught before it ships rather than after.
***
Ronn Torossian is founder of 5WPR, one of the leading PR agencies.
PR has more data than ever. The challenge now is turning it into clarity, context and better decisions.
For the first time, in-house teams and agencies have the opportunity not to just monitor narratives as they emerge, but to understand how they are forming.
Agencies are playing catch-up when it comes to drawing up the ground rules for AI. Getting up to speed starts with clearly defining how AI should be used.
The rapid growth of data centers across the United States is attracting attention and significant community concern. Aside from the heat of the debate, what does all that computing infrastructure actually do?
Proprietary opinion research is one of the most effective ways PR teams can strengthen brand visibility in AI search because LLMs prioritize original insights supported by hard data over recycled information.
