Escape Meridian

Max Roser · Nov 16, 2026

No matter how you ask, more Americans are negative than positive about AI

Markdown · JSON · Original source

Max Roser argues that, although survey wording affects how Americans report their views, a consistent pattern emerges across five recent probability-based surveys: more Americans view AI negatively than positively. In Pew’s survey, 52% said they were “more concerned than excited”, while 9% were “more excited than concerned”; Roser says the ratio had grown from about two to one when Pew first asked in 2021. He also points to concern across party lines and to respondents’ worries that future AI systems could threaten human survival. The pattern is not universal across questions: in an Ipsos survey, 50% called AI a good thing for them personally and 49% a bad thing. Roser’s argument is that the broader negative views matter for the AI industry’s relationship with the public: “If the industry wants to earn the public’s trust, it will have to take these widespread concerns much more seriously than it has so far.” The surveys describe reported views; the article does not establish what causes them.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

Set of stacked bar charts from five US surveys (from Pew, Gallup, Quinnipiac, Annenberg and Ipsos) where, however the question is asked, more Americans are neg…
Our World in Data · Set of stacked bar charts from five US surveys (from Pew, Gallup, Quinnipiac, Annenberg and Ipsos) where, however the question is asked, more Americans are neg… · Open original source

Original source

· Article

Source details

Original source date: Nov 16, 2026 (2026-11-16T00:00:00.000Z).

Date basis: Date stated by the publisher.

Added to Escape Meridian: 2026-11-16T00:16:53.094Z. First observed: 2026-11-16T00:06:22.671Z. Retrieved: 2026-11-16T00:15:54.346Z.

Referenced by

Recorded citations in public readings. A reference alone does not establish agreement or influence.

Some references could not be checked. Reload this page to try again.

Reference data (JSON)

Limits

The article reports survey responses and does not establish what causes Americans’ views or the difference between views about AI generally and views about its effects on respondents personally. The surveys were conducted at different times and used different questions; Roser notes that wording affects responses. Roser limited the analysis to surveys based on probability samples and excluded opt-in online polls.

Open question

What accounts for the difference between Americans’ generally more negative views of AI and the nearly even split when asked whether AI is good or bad for them personally?

Source-stated

Roser says survey wording affects reported views, but the surveys he reviewed all found more Americans negative about AI than positive.

the survey’s wording does matter. When a middle answer is offered, many people take it

Source-stated

In Pew’s survey, 52% were more concerned than excited about AI, compared with 9% more excited than concerned.

52% say they are “more concerned than excited”, and 9% are “more excited than concerned”

Source-stated

Roser reports that Ipsos respondents were nearly evenly divided on whether AI was good or bad for them personally.

50% said AI is a good thing “for you personally” and 49% a bad thing.

Source-stated

Roser argues that the AI industry should take widespread public concerns more seriously if it wants to earn public trust.

If the industry wants to earn the public’s trust, it will have to take these widespread concerns much more seriously than it has so far.

Source and contributor context

Max Roser · Article author

These are credits stated by the source; they do not establish a profile match or authorship of every passage.

Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Our World in Data

Transparent data interpretation and explanations of what headline metrics do not mean.

Editorial provenance

Version 1. Publication ID: e0727edd-e34c-439c-8699-9739c944769d.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T00:16:53.094Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: useful_perspective.

Observation: fc5038826f66f1add4b0f7dfeba818414f78a9e4a3fc9c469710d7e9339c8910. Source text hash: df886fd4db031af659e6613b166b30efb15efb54b183701cf983707ae776ae9b.

Revision and source-check history

Version 1 · 2026-11-16T00:16:53.094Z

No matter how you ask, more Americans are negative than positive about AI

Max Roser argues that, although survey wording affects how Americans report their views, a consistent pattern emerges across five recent probability-based surveys: more Americans view AI negatively than positively. In Pew’s survey, 52% said they were “more concerned than excited”, while 9% were “more excited than concerned”; Roser says the ratio had grown from about two to one when Pew first asked in 2021. He also points to concern across party lines and to respondents’ worries that future AI systems could threaten human survival. The pattern is not universal across questions: in an Ipsos survey, 50% called AI a good thing for them personally and 49% a bad thing. Roser’s argument is that the broader negative views matter for the AI industry’s relationship with the public: “If the industry wants to earn the public’s trust, it will have to take these widespread concerns much more seriously than it has so far.” The surveys describe reported views; the article does not establish what causes them.

Original source

· Article

Source details

Original source date: Nov 16, 2026 (2026-11-16T00:00:00.000Z).

Date basis: Date stated by the publisher.

Added to Escape Meridian: 2026-11-16T00:16:53.094Z. First observed: 2026-11-16T00:06:22.671Z. Retrieved: 2026-11-16T00:15:54.346Z.

Limits

The article reports survey responses and does not establish what causes Americans’ views or the difference between views about AI generally and views about its effects on respondents personally. The surveys were conducted at different times and used different questions; Roser notes that wording affects responses. Roser limited the analysis to surveys based on probability samples and excluded opt-in online polls.

Open question

What accounts for the difference between Americans’ generally more negative views of AI and the nearly even split when asked whether AI is good or bad for them personally?

Source-stated

Roser says survey wording affects reported views, but the surveys he reviewed all found more Americans negative about AI than positive.

the survey’s wording does matter. When a middle answer is offered, many people take it

Source-stated

In Pew’s survey, 52% were more concerned than excited about AI, compared with 9% more excited than concerned.

52% say they are “more concerned than excited”, and 9% are “more excited than concerned”

Source-stated

Roser reports that Ipsos respondents were nearly evenly divided on whether AI was good or bad for them personally.

50% said AI is a good thing “for you personally” and 49% a bad thing.

Source-stated

Roser argues that the AI industry should take widespread public concerns more seriously if it wants to earn public trust.

If the industry wants to earn the public’s trust, it will have to take these widespread concerns much more seriously than it has so far.

Source and contributor context

Max Roser · Article author

These are credits stated by the source; they do not establish a profile match or authorship of every passage.

Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Our World in Data

Transparent data interpretation and explanations of what headline metrics do not mean.

Editorial provenance

Version 1. Publication ID: e0727edd-e34c-439c-8699-9739c944769d.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T00:16:53.094Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: useful_perspective.

Observation: fc5038826f66f1add4b0f7dfeba818414f78a9e4a3fc9c469710d7e9339c8910. Source text hash: df886fd4db031af659e6613b166b30efb15efb54b183701cf983707ae776ae9b.

passed · 2026-11-16T00:16:53.094Z · anthropic/claude-opus-5.5

Publication: e0727edd-e34c-439c-8699-9739c944769d

attribution — pass: The draft correctly attributes the article and its argument to Max Roser at Our World in Data. Survey figures stay tied to Pew, Ipsos and the other named pollsters, and the closing recommendation is quoted as Roser's own view.

Max Roser (2026) - “No matter how you ask, more Americans are negative than positive about AI”

summary — pass: The summary leads with Roser's argument and includes the wording caveat, the Pew figures, the 2021-to-now shift, the cross-party finding and the survival-risk concern. It also keeps the Ipsos personal-effect exception and the author's call to the industry. Plain language makes it readable for non-specialists. The question about the general-versus-personal gap follows from the evidence. 'Nearly even split' fairly describes 50 versus 49.

What surprised me is how consistent the overall result is. Every survey finds that more Americans have a negative view of AI than a positive one.

date — pass: The publisher page and the archived citation both match the originalPublished date of 2026-11-16. No earlier first-publication date appears in the evidence.

(archived on November 16, 2026)

support — pass: Every claim is source-stated with exact figures: 52% versus 9% in Pew, the earlier two-to-one ratio, and 50% versus 49% in Ipsos on personal effects. The trust recommendation is quoted verbatim. The limitations accurately note that the article does not establish causes and that it was restricted to probability samples. The draft adds no unsupported benefits, costs or forecasts.

I only included surveys based on probability samples.

media — pass: The selected media is the verified direct link to the original Our World in Data article, which is the most direct way to reach the source. The verified chart image would also have been relevant, but the article link is appropriate.

Challenge or correct this reading

Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.