Only 9% of US Museum Visitors Accept AI in Exhibitions, AAM Survey Finds

By Tom Whitfield

Only 9 percent of US museum visitors accept the use of AI in exhibitions, according to the American Alliance of Museums' 2026 survey. The figure is the strongest signal yet that public anxiety about artificial intelligence now reaches into the gallery itself — and it lands just as several major institutions prepare to deploy the technology anyway.

What exactly do visitors reject?

The AAM survey breaks the unease into degrees. Nineteen percent of respondents said they were unbothered by AI being used for brainstorming, provided it did not extend to exhibition texts. Twenty-nine percent felt no reservations about AI writing emails sent by museum officials.

The hard line comes with exhibition content itself. Forty-three percent of respondents said they expected humans to "develop all content from museums, including exhibitions, emails…everything."

Read together, the numbers describe a public that tolerates AI at the administrative edges but rejects it at the interpretive core — the wall texts, labels and narratives that give exhibitions their authority.

AI is already arriving

AI is not yet widely used to produce exhibitions. That is about to change, at least in a handful of cases. A New York Times report published the first weekend of November 2026 revealed that several forthcoming shows will incorporate AI, though not always in the form of wall texts.

The most concrete example: the Smithsonian's National Museum of Asian Art will involve AI in a virtual reality recreation of James McNeill Whistler's Peacock Room, part of its version of a Tate-organized retrospective for the artist.

Note what the survey does not say. It measures attitudes, not behavior — whether visitors who claim to reject AI would actually stay away from an AI-assisted exhibition is a question the data cannot answer.

A second problem: not whether, but how often

The survey also addressed museum attendance, a perennial concern since the pandemic limited visitorship in 2020. Many institutions have still not recovered. The AAM did not explicitly measure annual visitor counts. Instead, it flagged a different problem: repeat visitation.

"In the six years since the pandemic began, this repeat visitation has leveled off into a lower 'new normal' that seems to reflect enduring changes in behavior," the survey said. "In other words, it's not whether people are visiting museums that's causing attendance shortfalls; it's how often."

That reframing matters for budget planning. If the problem is frequency rather than reach, the remedy is programming that rewards return visits — not marketing that chases first-timers.

Who is actually visiting?

The survey also challenges long-standing assumptions about the museum audience, which institutions previously saw as "whiter, older, and more educated than Americans on average."

The current data tells a different story:

• Only 47 percent of museum visitors hold a college degree.
• Only 35 percent are white (not Hispanic or Latine).
• Only 26 percent are 60 or older.

"White respondents are the most likely to report 'never' visiting museums (while Asians and Asian Americans are the most likely to report visiting), and adults over 60 are the least likely of any age/life stage group to have visited in the past year (while parents or guardians of minor children are the most likely)," the survey said.

What does this change for museums?

The public appears to want museums to remain inclusive, even as the current Trump administration seeks to change that. "While younger people and liberals show the most inclusive attitudes, the majority of all age segments and the majority of conservatives still fall on the side of inclusion," the AAM survey said.

That leaves institutions squeezed between a public that expects human-made content and broader inclusion, an administration pushing in another direction, and a handful of upcoming shows — starting with the Smithsonian's Whistler retrospective — that will test whether the 9 percent figure reflects real resistance or merely stated preference.