Corporate Reputation Risk Rises as AI Engines Surface Decades-Old Scandals
A recent study conducted by Sam Michelson of Five Blocks and shared via PR Daily reveals that AI platforms frequently resurface historical corporate crises. The research found that nearly half of the serious negatives mentioned in AI-generated company descriptions concern events from 2020 or earlier, including incidents dating back to the late 1980s.
Context from Industry Pulse
The PR Daily study highlights a fundamental shift in crisis management where the news cycle no longer naturally buries past failures. While traditional media may move on, AI engines treat historical data as current components of a company's profile if recent resolutions are not explicitly documented in trusted locations. The data shows that while recent events are less likely to be mentioned, they carry a higher probability of being negative when they do appear. This creates a double-sided challenge for brands: they must manage immediate negative sentiment for new events while simultaneously battling the long-tail digital persistence of legacy scandals that AI tools continue to prioritize in their summaries.
"If the resolution was never written down in a place the engines trust, the old coverage is the most complete account in the file. So that’s what gets repeated."
Research published by PR Daily indicates that artificial intelligence tools lack the expiration dates found in traditional credit reporting, leading to the persistent surfacing of ancient corporate controversies. A study of Fortune 100 companies utilizing the AIQ platform examined approximately 25,000 responses across eight different engines. The results showed that 43 percent of serious negative mentions involved lawsuits, safety issues, or scandals occurring at least four years ago. Some engines still highlight the 1989 Exxon Valdez spill or 2009 bankruptcies, regardless of subsequent leadership changes or legal settlements.
The inconsistencies between different AI models are significant. According to PR Daily, about one-third of identified negative points appeared in only a single engine among the eight tested. Even products from the same developer, such as Google AI Mode and AI Overview, showed a massive disparity in how often they raised serious corporate issues. This lack of consensus suggests that monitoring a single platform is insufficient for internal communications teams trying to manage a digital reputation.
Source attribution plays a critical role in how these engines formulate their answers. PR Daily reports that company websites are the most cited sources, appearing in 82 percent of the collected answers. Wikipedia follows as the most prominent external source, though its influence varies wildly between models. For instance, Google AI Mode cited Wikipedia in nearly all responses, while ChatGPT rarely did so. Other frequently referenced platforms include Yahoo Finance and Britannica.
To mitigate the impact of outdated information, PR Daily suggests that companies must actively update their own digital records. Because engines prioritize available documentation to complete a narrative, firms are encouraged to post factual resolutions of past crises on their own governance or history pages. This provides the necessary data for AI to balance old news with current realities. Communications teams are advised to run regular monthly or quarterly audits across multiple engines to identify emerging or resurfacing reputational threats before they escalate.
The Industry Pulse Bottom Line
Brands must treat their own websites and Wikipedia talk pages as primary data feeds for AI. Without factual, accessible updates on crisis resolutions, AI models will continue to rely on archived negative reporting indefinitely.