Reputation Radar
PR, corporate and crisis communications, sourced weekly.
Five stories from the past week, with the takeaway that matters for reputation leaders. This week: Meta heads to federal trial over claims it designed Facebook and Instagram to addict kids, AI resurfaces a promotion that ended a year ago and sends real customers to a real door, and two writers make the case for why structure, not just substance, decides whether AI cites you at all.
Meta Goes on Trial Over Claims It Designed Apps to Addict Kids
NBC News · Aug. 18, 2026
A federal trial got underway this week in Oakland, where four states leading a 29-state coalition, California, Colorado, Kentucky and New Jersey, argue Meta knowingly designed Facebook and Instagram to addict young users. Testimony began Tuesday, including from a former Meta safety researcher who told the court the company’s own definition of “problematic use” undercounted what researchers would otherwise call addiction. The case is expected to run about six weeks, and even the numbers at stake are disputed: reporting has cited a roughly $200 billion damages estimate, though Meta’s own court filings put the figure states are actually seeking closer to $1.4 trillion. Whichever number holds up, the case tests something bigger than one company’s messaging: whether a product’s design choices, not just what a brand says about itself, are now the central battleground for reputation.
AI Just Created a New Kind of Crisis: Recycled Truth
PRNEWS · Aug. 17, 2026
A restaurant’s 2025 Teacher Appreciation and Nurses Week promotion, which it chose not to repeat this year, got resurfaced by AI search tools as if it were still active, sending customers to the door expecting a deal that no longer existed. PRNEWS editorial lead Kristen Kauffman calls this a new risk category: not fabrication, but legitimate information stripped of its timestamp and published without verification. Her fix involves three protocols most crisis matrices don’t yet have: auditing where old campaign content still lives online, building expiration signals into promotional content, and having a media remediation plan ready before it happens. The uncomfortable takeaway is that a brand doesn’t need to say anything wrong to end up managing a crisis it didn’t create.
AI Can Help You Find a PR Firm. It Can’t Tell You Who to Trust.
O’Dwyer’s · Aug. 17, 2026
WOC | Signal managing director Larry Smalheiser argues AI has become a real front door for agency discovery, prospects now find firms through generative search the way they used to find them through referrals. But he draws a sharp line between discovery and discernment: AI can surface credentials and sector experience, but it can’t tell a client how an agency will react when a strategy stops working or a board member pushes back. His argument is that case studies and awards help, but trust still gets decided the old way, through direct conversation and judgment under pressure. For an industry increasingly worried about AI visibility, it’s a reminder that being findable and being trustworthy are two different problems.
Want AI to Cite You? Give It a Map.
PRsay · Aug. 17, 2026
PRSA’s John Elsasser makes a practical case that structure, not just quality, decides whether AI tools pick up your content. Clear headings, organized sections and scannable formatting help large language models locate and extract exactly the piece they need, while dense, unstructured writing gets passed over even when the substance is strong. The advice lines up with what other AI-visibility researchers are finding: machines don’t reward good writing the way human editors do, they reward writing that’s easy to parse. For communicators already thinking about earned media and AI citations, it’s a reminder that how something is written now matters almost as much as what it says.
This Year’s Biggest Reputation Comebacks Share One Thing in Common
Reputation Sciences · August 2026
New reputation rankings show UnitedHealth Group, Tesla and Southwest posted the largest reputation gains of 2026, all three after working through very public backlash in recent years. The rankings also reveal a split inside AI itself: Nvidia’s reputation climbed on the strength of powering the AI boom, while OpenAI dropped 13 spots despite explosive growth in brand value. The pattern suggests the market currently trusts the companies building AI’s infrastructure more than the companies building AI’s products. For crisis-tested brands, the larger lesson is that a bad year doesn’t have to be a permanent one, provided the recovery is visible and sustained, not just declared.
I pull five stories like this every Friday: one crisis case study, one industry shift, one signal on how AI is reshaping trust in this profession. If your team is navigating one of these in real time, reach out.
Follow on LinkedInEvery Friday · generodriguezmiller.com