How the National Gallery of Art is Using AI to Unlock New Insights into Art and Pain Managment

The National Gallery of Art in Washington, D.C., is pioneering new ways of using artificial intelligence (AI) to broaden how collections are studied and applied across disciplines. With its long-standing commitment to digitization, open access, and collaboration, the National Gallery has become a leading example of how museums can contribute to cutting-edge research beyond the traditional boundaries of art history.

A Foundation in Digitization and Access

Since opening in 1941, the National Gallery has welcomed millions of visitors to explore its world-class collection. Today, it houses more than 160,000 works of art, ranging from masterpieces of the Renaissance to modern highlights. But it is the museum’s digital programme that has positioned it at the forefront of AI-driven research.

The institution began large-scale digitization in 2004, adopting FADGI standards and introducing innovative tools such as a robotic easel for consistent image stitching. Its digital asset management system now handles more than 1.6 million assets, all fully integrated into the public-facing website. More than 61,000 high-resolution images are freely available for download, reflecting the museum’s strong open access policy.

“AI models require structured, consistent, high-quality data to function accurately and reliably,” explained Justyna Badach, Head of Imaging at the National Gallery of Art. “Because of our long-standing digitization efforts and metadata standards, our digital collections are ideally suited as controlled datasets for AI research.”

Through partnerships such as PHAROS, a consortium of 13 photo archives that will release 2.6 million images for public use in 2026, the museum is contributing to some of the most significant open resources for art and AI research worldwide.

From Art to Neuroscience

One of the most innovative applications of these resources emerged through a collaboration with neuroscience researchers at McGill University. Hannah Derue, a PhD candidate specialising in chronic pain, founded PAin+, a software platform that combines arts-based mindfulness, neuroscience, and machine learning to provide drug-free interventions for people living with pain.

The idea grew from her encounter with the National Gallery’s imaging facilities. “I realised we were asking the same kinds of questions,” Derue recalled. “Conservators were using imaging to reveal hidden layers of paintings, and in neuroscience we’re using similar tools to uncover hidden patterns of pain.”

The National Gallery’s open access collection became the foundation for training PAin+’s AI models. By analysing thousands of artworks, researchers explored how motifs of pain are expressed visually and textually, with the aim of helping patients better articulate their own experiences.

Giving Form to the Experience of Pain

Communication remains one of the greatest challenges in pain care. Pain is highly personal and culturally shaped, making it difficult to standardise or diagnose. PAin+ addresses this gap by providing users with visual and textual tools to record and reflect on their pain, creating outputs that can be shared with healthcare providers.

Working with programmer Patrick Chen, the team also developed a natural language processor trained on anonymised real-world pain reports. This “pain lexicon” doesn’t just recognise words; it interprets the context and emotional weight behind them. Insights are presented in personalised dashboards, enabling patients and clinicians to track patterns, detect flare-ups, and assess treatment responses.

“Think of PAin+ as a pain journal that listens, learns, and adapts,” Chen explained. “It translates personal stories of pain into data that can improve communication and ultimately care.”

Art and Science in Partnership

For Derue, the collaboration shows what can happen when disciplines connect. “By combining science, art, and machine learning, we’re building a more complete and compassionate way of understanding pain,” she said.

For the National Gallery of Art, it is another example of how open access and digitization can extend the impact of collections. As Badach reflected, “We aim to increase access to art, but that  isn’t just about images on a website. It’s about enabling knowledge, creativity, and innovation in ways we couldn’t imagine on our own.”

–

Justyna Badach from the National Gallery of Art and Hannah Derue from the Alan Edwards Centre for Pain Research at McGill University spoke at the Digital Collections Summit in September 2025.