AI search engines give outdated negative content indefinite half-life. Traditional reputation management does not work anymore. Most businesses have no idea they are exposed.
TL;DR: AI search engines cite sources based on authority signals, not recency. A negative article from a decade ago that faded from Google rankings can be resurrected by AI Overviews or ChatGPT as a primary source about your business. Traditional reputation management, publishing fresher content to push negative results down, breaks because AI accesses and cites the original source directly. The half-life of bad press used to be measured in search ranking decay. Now it is infinite.
A Grocery Chain, a Resolved Crisis, and an AI That Would Not Forget
There is a grocery chain whose mid-2010s customer service issue was fully resolved years ago. The company fixed the problem. Customers moved on. The negative article faded from Google search, buried under years of newer content. No one searching Google today would find it near page one.
Google AI Overviews found it anyway and cited it as an authoritative source about the business.
The issue was historical. The resolution was complete. The article was old. None of that mattered. The AI selected it based on authority signals, outlet credibility, citation history, rather than recency or current accuracy. The company solved its problem a decade ago. For AI search purposes, the problem is current. Anyone who asks an AI about this grocery chain might be told about an issue resolved before some of their customers were born.
This is not an edge case. If your business has ever had a public crisis, a negative article, a regulatory action, a lawsuit, the same mechanism is applying to you right now. You just do not know it yet.
Why Traditional Reputation Management Breaks
For fifteen years, online reputation management followed a reliable playbook. Publish fresher content. Optimize it for search. Push negative results down in rankings. Control the first page of Google and you controlled what people saw.
AI search breaks this playbook in three ways. First, AI accesses and cites original sources directly. It does not care what ranks on page one of Google. A negative article on page ten can still be the source an AI Overview chooses. Second, suppression content rarely gets cited. The kind of thin, keyword-optimized content built for ORM is exactly what AI systems deprioritize. Third, AI may cite both sides, stitching together the original negative article and your suppression content into something that sounds balanced. To a reader, it sounds like an ongoing issue. The suppression content did not suppress anything. It confirmed there was something to suppress.
Think about a library that keeps every newspaper ever printed. Someone asks the archives for information about your business. The most cited, most discussed article from 2012 comes right to the top, regardless of whether it is still accurate. That is how AI retrieval works. Bad press used to have a half-life. Now it is infinite.
What Makes It Worse
AI search is not one thing. ChatGPT routes your question through different retrieval pipelines you never see. One pipeline searches established publishers. Another searches commercial databases. Another searches the open web. When the pipeline changes, the sources the AI cites change about 45% of the time. The same question asked twice can cite entirely different sources about your business.
For reputation monitoring, this is a nightmare. Your brand might look fine on one query because the pipeline surfaced recent positive coverage. On the next, it surfaces a decade-old complaint. You cannot see the difference. You cannot reproduce the bad result. You just have a customer who saw something negative and a brand that has no idea it happened.
The New Playbook
We have seen this pattern across industries. Product recalls, lawsuits, regulatory actions. The older the coverage, the more likely the business assumes it is handled, and the more exposed they are. Professional services firms are especially vulnerable. Law firms, financial advisors, medical practices. The kind of businesses where trust is the product and a single negative citation can disqualify you from the shortlist an AI assembles before a buyer ever contacts you.
The old approach is broken. The new one starts with earned media. Muck Rack's analysis of 25 million AI-cited links found that earned media drives 84% of AI citations. The publications and journalists who cover your industry are the same sources AI systems trust. Every legitimate placement is an AI citation asset. Speed matters because every week a negative source goes unchallenged, it gains citation history that makes it harder to displace. And monitoring matters because you cannot fix what you cannot see. Most businesses are not checking what ChatGPT, Perplexity, or AI Overviews say about them.
Here is what makes this moment unusual. AI-era reputation management is a real vulnerability that almost no one is prepared for. The businesses that move first, auditing their AI visibility and building citation-worthy earned media, will have a structural advantage. Their competitors will not realize anything is wrong until a customer asks why an AI describes their company in the past tense.
Your reputation in AI search is not what you published last week. It is what the most authoritative source about your business published, whenever that was. The only question is whether that source is you, or someone who wrote about you a decade ago.