Fearing the Digital Future? Get AI Literate!

2026 GHHN Annual Conference & Awards for Excellence
What We’re Afraid Of: Challenges, Changes, and the Courage to Face Them

Tuesday, October 27, 2026
Hudson River Museum
Yonkers, NY
9:00 AM – 4:45 PM

The conference asks, “What are we afraid of?” For some of us, the fear might be:

That does sound scary to this longtime museum educator. Is there anything we can do to stop this erasure from happening? 

Answer: YES! Get AI literate. We cannot adequately address or protect museums from the invasion of artificially intelligent systems unless we understand some basics about how AI works, what it requires, and, most importantly, where you can use it for your own advantage. Here are just a few thoughts to inspire your AI literacy journey:

I. Artificial Intelligence is already inside your galleries

Whether you —the museum staff members—use AI or not, your public will be using it, either intentionally or unintentionally through existing apps that now integrate AI. For example:

Is translating materials into more languages on your wish list? 

Feel free to remove it now because guess what your visitors are doing today? Pointing their phone cameras at your labels and getting back translated ones via AI tools such as Google Translate.

Are you worried that AI will mistranslate information? That is possible but language facility is one aspect of Generative AI that has improved the fastest over the last 3 years. Translation systems are regularly tested and now score nearly on par with professional translators, especially for the most commonly spoken languages. 

Are visitors using AI to understand other aspects of our museum? 

Yes, definitely. They point their phone cameras at anything in your building and collections. Google Lens is a free tool that quickly identifies, explains, and/or transforms images collected through the phone camera.  Moreover Google AI, now called Gemini, remembers what is identified in the user’s profile. Don’t be surprised if Gemini sends you “out of the blue” suggestions about where you can see nightscapes like the ones you were asking Google lens about.

II.  Museums are using AI to manage long-standing operational challenges

Is AI good at improving visitors’ experiences and engagement?

We know visitors enjoy variety and surprise in their museum experience: playing a game, conversing with a historical figure, or offering a family activity. Can we ever generate enough of these experiences to satisfy every visitor? Seems unlikely in the analog world. 

Consider these examples of  museums experimenting with AI to improve visitors’ experiences: 

A unique sculpture depicting a stylized human face with exaggerated features, made from a textured, light-colored material, topped with a small black figure resembling a seated individual.
Screenshot of an app titled 'The Effects of Alcohol' featuring a user interface that includes a question about eye-drawing features and options to send messages.
A carved bone figure resembling a human head with intricate textures and a surprised expression, standing on a curved base.

Front and side view of a sculpture entitled “The Effects of Alcohol” made of bone and stone, depicting a woman with a bottle of alcohol lodged in her back created by Inuit artist Kavavow Pee in the collection of the Museum of Fine Arts Quebec. Center panel shows AI screen asking a question about the visitor’s perception of the object.

  • The Cleveland Museum of Art made an AI tool that can quickly and authoritatively generate visual descriptions using images and documentation of art. This is a particularly important feature for visitors who cannot see well, as well as websites whose image alt tags are not in compliance with current accessibility guidelines. Also, It will make it easier for Cleveland’s collection to be found by search engines, especially those powered by AI. 

Could these AI products generate inaccurate or misleading information?

Not likely. Why? These AI systems are trained with museum-supplied materials & human intelligence. Compare this strategy to ChatGPT which is trained on a universe of disparate facts that the LLM’s must make their own sense of. The AI products described above are made from very limited sets of authorized information and are trained to perfect their understanding of them. These RAG (Retrieval-Augmented Generation) systems are known for their high-level of accuracy in managing specialized data sets. 

Have FOMO? Don’t even think about it. Let the big museums with dedicated budgets for experimentation be the innovators for now, a likely outcome is that these AI-based tools will soon become available and adaptable for everyone’s use, and probably for a low or no cost.

  III. What can I do now? Build your and your team’s AI literacy!

  1. Start by forming a regular (monthly lunch time?) study group with a few staff members that share your AI curiosity. Each of you could learn more about AI from different perspectives, and then share your findings in subsequent discussions, kind of like a book club but the topic is AI. 
  • Grow a set of ideas and resource links or whatever best represents the group’s findings. Share this internally to see if other staff become inspired. 
  • Learning can be as easy as starting with the New York Public Library’s Tech Connect program. Their staff is AI knowledgeable and teach AI literacy and practice. NYPL online and onsite programs are free.
  1. “What do YOU use AI for?” might be an interesting “ice breaker” question to pose before/during/after a public tour. Learning how your visitors use AI may inspire new thinking about how your museum might use AI. 
  • Keeping conversations open and shared helps museum staff become more confident in their command of AI.
  • Encouraging visitors to teach us (for a change!) about AI shows that museums can be good listeners too.
  1. Create AI Guidelines for staff use. As you progress in your AI literacy journey, you will encounter aspects of AI that your staff do not want and/or should not use. Keep a list of these and explain your reasons why. Eventually, this list could become the basis of your museum’s   official guideline for AI use. See for example the Cleveland Museum of Art’s “Using AI Ethically and Responsibly at the Cleveland Museum of Art” in the resource section below.

Good luck, enjoy the AI learning ride, and please reach out to us if you need any help. 

Howes Studio

Deborah & Sophia Howes


Resources

IEEE Spectrum: “ChatGPT’s Translation Skills Parallel Those of Most Human Translators” — A readable account of a peer-reviewed study comparing AI translations with translators at different experience levels. The study found that results varied: AI was comparable to less-experienced translators in its tests, while senior translators did better. The article’s headline is broader than those specific tests warrant. 

W3C’s Images Tutorial — The best all-around guide. It explains what to do with informative, decorative, functional, and complex images, such as charts. The central idea: describe an image’s purpose and important information, not just its appearance. 

The Cleveland Museum of Art used Gemini to help create descriptions for more than 68,000 collection images. The descriptions follow a style guide developed with accessibility experts, are labeled as AI-generated when applicable, and serve as screen-reader-accessible alt text on the website. Descriptions that fall below the museum’s automated quality threshold are flagged for human review. “Making Art Accessible to All: Visual Descriptions for Every Artwork in the Collection”. 

Crystal Bridges, AR is offering an AI art companion to visitors who seek a personalized storytelling experience. Visitors open the “Rosie” webApp (no downloading) on their smartphone and show an artwork to the screen to begin a conversation. It is voice-enabled by design so that visitors can keep their eyes on the art while they talk and listen; the goal is to support close looking. The foundational training consists of exclusively museum generated and approved content. 

The Cleveland Museum of Art created an online game “Share your View” that allows you to upload images of your own for the AI tool to find something similar in the collection. It’s like having a visual conversation with an AI engine which shows you how the computer “sees” art thru decoding your picture and using metadata in the collection database to find matches.

As a cultural leader, the CMA believes audiences expect museums not only to adopt new technologies thoughtfully but also to help establish standards that protect artists, respect intellectual property, and preserve the value of human creativity. Our use of AI is governed by an internal institutional policy that emphasizes the following:

  • Alignment with the CMA’s mission and values
  • Respect for artists, cultures, and intellectual property
  • Avoidance of harmful bias or misrepresentation
  • Human review for all public‑facing outputs
  • Clear accountability for decisions influenced by AI

The CMA does not allow AI systems to make unsupervised decisions that affect individuals and we actively test and evaluate systems to mitigate bias and unintended consequences. We view AI to be a tool that must be wielded responsibly, which can be used creatively, but must not diminish artistic authenticity.

The CMA takes data protection seriously. We do not share confidential, proprietary, or sensitive information with public AI tools, and we carefully evaluate vendors and systems before approving AI technologies for use. Images of artworks and other copyrighted materials are handled with particular care, and AI systems are not permitted to retain or train on protected content.

Also, see “Day of AI” from MIT RAISE (Responsible AI for Social Empowerment and Education” for curriculum resources on Ai literacy.