Journalist documents her frustrating quest to fix an error in the Knowledge Graph panel appearing on Google searches for her name, which claimed she was dead
I'm not dead yet. — But try telling that to Google. — For much of the last week, I have been trying to persuade the world's … See also Mediagazer
Context & Ripple Effects
The piece lands two months after BuzzFeed documented how Facebook, Google, and Twitter were failing at their pledged accuracy during major crises — this time the failure is personal rather than crisis-scale: Google's Knowledge Graph panel told searchers a living New York Times journalist was dead, and her account is a first-person audit of how hard it is to correct an automated 'fact' about yourself.
It also arrives in a year of friction between Google and the press — weeks earlier, another reporter described pressure from Google and her editors to retract a 2011 Forbes post criticizing the company's search practices. Seven years on, the arc continues: by 2024, AI-generated obituary sites were listing living people as deceased across Google search, turning the Knowledge Graph's one-off error into a spam category.
First-order effects
- The journalist spends days inside Google's correction process with no guaranteed fix, while every searcher of her name sees a false death claim attributed to a source that looks authoritative.
- Google faces public documentation of its error-correction path being too slow for a factual claim about a named, identifiable person.
Second-order effects
- Publishers and public figures learn that Knowledge Graph errors carry the panel's authority but none of its accountability, pushing them toward public shaming — press coverage — as the only working escalation channel.
- The gap between authoritative presentation and correctable content creates the opening later exploited by obituary-spam sites, which game the same search surfaces for clicks.
Third-order effects
- If automated fact panels keep outrunning correction mechanisms, the pressure shifts toward regulators and courts treating machine-assembled claims about individuals as a liability question rather than a customer-service ticket.
- Search results about people drift from curated reference toward unvetted aggregation, eroding the trust that makes panels valuable to Google in the first place.
The trend: Search engines are presenting auto-aggregated facts about individuals faster than they can correct them, a gap running from 2017's Knowledge Graph errors to today's AI-generated obituary spam.