Artificial intelligence is appearing in all corners of museums. Staff are using it to organize collections, identify artifacts that need attention, manage their membership tools, and observe how visitors move through galleries. These new mechanics are starting to support curators and educators in ways that make everyday work more manageable.
Recent research on human-centric AI shows that museums are exploring these possibilities with intention. Rather than following trends, institutions are folding AI into existing practices, asking how it can improve daily tasks and strengthen their work with the public. What we see is careful progress and a growing sense of what’s possible.
At Cuseum, we examined the latest research on human-centric AI. What we found is a field at a full of optimism, but still looking for ways in which to better incorporate this tech into its already functioning systems.
The Reality: What AI Is and Isn’t in Museums
Since 2021, AI has appeared in a growing number of museum activities. Teams have started using it to implement some interesting innovations like: catalog artworks, detect patterns in visitor movement, and monitor the condition of fragile pieces. In some institutions, AI has even suggested new ways to present stories or help staff answer questions more quickly.
Rather than relying on one size fits all solutions, museums are choosing the AI tools that make sense for their own collections and communities. This thoughtful approach creates space for creative projects and prepares institutions to adapt as technology continues to evolve.
Museum leaders are consistently pointing out a fundamental truth: AI is not about replacing people, but supporting them.
“This [AI] is a fascinating development, but it can be tricky to implement,” admits one digital education specialist (Derda & Predescu, 2025). Others see AI as a “helping hand” for translation, accessibility, or making collections more searchable, but no one is clamoring to surrender the curatorial vision to an algorithm.
What is happening? Let's get to the bottom of this:
Some institutions are piloting AI for image recognition, sentiment analysis, and visitor flow optimization—“practical, limited-scope tools that can enhance what we already do, not upend it”.
The biggest headline use cases? Automated transcriptions, multilingual audio guides, and personalized content recommendations, especially for larger museums with sufficient resources.
Many teams still rely on external partners and design studios for anything more advanced than a chatbot or a basic analytics dashboard.
Some smaller museums have chosen to hold back from adopting AI, often pointing to cost and complexity as challenges. Even so, AI is being built in many forms, and a growing number of accessible tools are beginning to catch the attention of institutions of all sizes. As options continue to expand, there is a sense of optimism that more museums will find the right fit for their needs, regardless of their size or resources.
Attitudes: Optimism Meets Expectations
When asked about AI, most museum professionals express a mix of curiosity and anticipation. Concerns about job loss or bias sometimes appear, especially with so much dramatic coverage in the media. Yet for many, the real questions are rooted in purpose. They want to make sure new technology supports the heart of their mission and preserves what makes museums unique. This ongoing conversation is shaping a thoughtful path forward, as teams look for ways AI can help without losing sight of their values.
As one head curator put it:
“Maybe that’s the [core] question—how important do you make artificial intelligence as a museum? Is it a tool? Or is it the way that you interact with visitors? ... I’m really excited about the answers we can come up with.” (Derda & Predescu)
Several recurring themes came up in our review of the field:
AI can increase access (translation, content, wayfinding) and make certain operations more efficient, but most museums don't want AI to define visitor experience or take curatorial control.
Staff take into consideration ethical challenges—privacy, bias, data misuse, and the risk of eroding public trust.
There’s some concerns that AI could “distract” from core experiences, overcomplicate simple pleasures, or create barriers between visitors and the art itself:
The Data: Where Are We in 2025?
Let’s ground this in numbers:
68% of European museums now have permanent collections online (Netherlands Museums Association, 2022).
True AI adoption in visitor-facing contexts is still being defined. It is mostly pilots, not platform-wide solutions, that have been incorporated.
According to the Benchmark Report on Cultural Website Performance, 40% of cultural organizations had a mobile-optimized site by 2023, a precondition for AI-powered personalization and accessibility (Benchmark Report, 2023).
Most professionals surveyed rated their own knowledge of AI as “still learning”, which opens the gate for multiple possibilities, and becomes a major factor in careful adoption.
This is a healthy position to be in.
AI is a highly promising utility but we don’t expect it to do everything. And in a sector built on trust, educational value, and slow thinking, that opinion might be a superpower
Why Human-Centric AI (Still) Matters
What does human-centric mean in practice?
In short: every step toward automation, optimization, or AI-powered convenience must be filtered through the museum’s core values and the visitor’s needs.
AI is useful when it helps real people, staff or public, access collections, find meaning, or participate in the life of the institution. But it can not produce impactful results when it solely chases novelty. As Graham Black (2020) and others have argued, museums are “bastions of reliable knowledge” in a chaotic information landscape. Losing that title is not an option.
The good news? Most practitioners agree. Human oversight, ethical frameworks, and transparency are at the heart of every serious AI conversation in museums. These tools are only as valuable as the mission they serve.
What Comes Next: Laying the Foundations
So, where does this leave us? Not necessarily in a sci-fi future, but in a landscape of methodical, considered, incremental experimentation.
AI is a supportive tool, not a replacement.
Visitor-centered design, ethical oversight, and human curation are a good starting point.
Pilot programs are popular, but full-scale deployments will probably be employed at larger scale in the future.
This is a sector with its eyes open: willing to experiment, but determined to keep people at the center of the story.
In the next article, we’ll move from the big picture to the practical: Six Principles Every Museum Should Use for Visitor-Focused AI. We’ll break down how human-centric frameworks (like the HC-AIM) can turn cautious optimism into real value, without losing what makes your institution unique.