# Chat, what’s the public’s favorite AI lab? > Source: > Published: 2026-08-03 10:01:56+00:00 # Chat, what’s the public’s favorite AI lab? ### OpenAI is winning the war for public opinion. Today’s leading AI labs claim to be determining the course of humanity’s future. Those claims can seem grandiose to the average chatbot user, who may view programs like ChatGPT as advanced forms of Google search. But recent dust-ups with the U.S. government show that the stakes are already getting higher. In February, Anthropic and the Defense Department feuded over whether the military should have access to an unfettered version of Claude that could conduct mass surveillance of Americans and operate weapons autonomously[1](#footnote-1). The immediate standoff ended when a competitor, OpenAI, swooped in to offer the department access to ChatGPT on more favorable terms. This episode encapsulates the reputations that the two companies have developed among those in the know: Anthropic is considered the safety-conscious lab, with a thoughtful CEO who pens essays about the [promise](https://darioamodei.com/essay/machines-of-loving-grace) and [peril](https://darioamodei.com/essay/the-adolescence-of-technology) of the technology. OpenAI is the unscrupulous competitor whose CEO was once ousted by the company’s board for [“not being consistently candid in his communications.”](https://www.cbsnews.com/news/openai-sam-altman-returning-as-ceo-will-build-microsoft-partnership/) This perception, however dominant in elite discourse, appears not to have filtered through to the general public *at all*. Six months ago, very few people had even heard of Anthropic, although it has essentially [cast itself](https://www.wired.com/story/anthropic-thinks-ai-can-only-be-safe-under-its-control/?_sp=a7a148ef-f17a-490a-9e2b-dd1ec3e37ea4.1782918398708) as humanity’s protector. A December 2025 [poll](https://www.searchlightinstitute.org/research/americans-have-mixed-views-of-ai-and-an-appetite-for-regulation/) by the Searchlight Institute found that 81% of respondents had no opinion on Anthropic. Another [analysis](https://x.com/davidshor/status/2001826280011137229) around that time by Democratic data scientist David Shor suggested that the true proportion of people who hadn’t heard of Anthropic could be as high as 95%.[2](#footnote-2) While that number seems to have fallen dramatically over the past few months — now that Anthropic has overtaken OpenAI in terms of annualized [revenue](https://www.saastr.com/anthropic-just-passed-openai-in-revenue-while-spending-4x-less-to-train-their-models/), run [funny](https://x.com/sama/status/2019139174339928189?lang=en) Super Bowl ads, and clashed with the Trump administration — a lot of people *still* don’t know what Anthropic is. *The Argument’s *poll of 3,008 registered voters fielded between May 29 and June 3, 2026, found that 42% of respondents had not heard of Anthropic and 38% did not know xAI. OpenAI led the pack in name recognition, with just 13% of respondents saying that they had not heard of the company — a whopping 29 points better than Anthropic. All told, this means that Anthropic gained somewhere between 39 and 53 percentage points in name recognition over the past six months. But increased public awareness of Anthropic hasn’t necessarily translated into favorability.[3](#footnote-3) When we asked respondents whether they had a favorable or unfavorable view of each AI company, OpenAI led favorability by a healthy margin. Among people who knew of each company, OpenAI had a net favorability rating of +11 points, while Anthropic had a net favorability rating of only +3. Whether or not you think Anthropic’s self-styled heroism is [genuine](https://www.anthropic.com/news/position-open-weights-models), you might expect it to have *some* discernible effect on public opinion. After all, while few people have read CEO Dario Amodei’s musings, millions watched Anthropic take a shot at OpenAI in Superbowl ads or heard about its choice to stand up to the Trump administration. You could even imagine a world in which the company’s earnest approach has the opposite of its intended effect: The clash with the Trump administration, say, could politically polarize views of the company, leaving MAGA supporters with a bad taste in their mouths. But there seems to be no good evidence that any of this has happened yet. In fact, it seems that public approval of individual AI labs isn’t driven by much of *anything* they say or do, outside of delivering a useful consumer product. ### OpenAI isn’t Kleenex, it’s Amazon My starting hypothesis was that OpenAI’s lead in net favorability was due to its first-mover advantage. Maybe the public had come to associate AI with OpenAI above all other companies, the way Coke stands in for soda or Kleenex for tissues. This led me to suspect that frequent AI users, being more aware of the various companies’ brands, would look different from those who barely used AI at all. ## Keep reading with a 7-day free trial Subscribe to The Argument to keep reading this post and get 7 days of free access to the full post archives.