Art has long been a powerful way for communities to express and share their stories. Whether through a song, painting or movie, art helps individuals and communities make sense of their emotions, their experiences, their histories and their futures, and to convey this across generations and into the wider world.
Art can also be used to help us when there are problems. For example, artistic expression is an evidence-based form of therapy for people with mental health problems and is a key part of effective mental health treatment systems. In South Africa, storyboarding workshops with young women living with perinatal HIV revealed ways to improve treatment adherence. These stories, later shared through a documentary, reached policymakers, educators and healthcare workers. On a broader scale, art can also foster community healing. For instance, TAIBU Community Health Centre in Toronto used art to promote mental health in Black communities by encouraging reflection on Afrocentric values. These deeply personal processes thrive in safe spaces.
The advent of artificial intelligence (AI) is now challenging the power of art for healing, empowerment and better health, however. By training on copyrighted content, AI tools have sparked legal disputes across the U.S., U.K. and Canada. But these lawsuits only hint at a deeper issue tied to health equity.
The key question is: Can computer-generated formulas genuinely capture the heart of artistic expression in equity-seeking communities and individuals?
Dr. Edsger Dijkstra claimed, “A picture may be worth a thousand words, but a formula is worth a thousand pictures.” Written in 1996, this phrase foreshadowed the algorithms that now analyze vast and often not diverse datasets to create new forms of art.
AI-powered art therapy is gaining momentum as a way for healthcare patients to express themselves. On the other hand, challenges with AI may arise as computer vision models are being developed to interpret emotions through art or to create narratives based on vulnerable patients’ works.
Without regulation, these tools can strip original images of their context and meaning, then use them to generate anything. This threatens to erode trust in using art for healthcare, particularly in communities already skeptical of AI and the healthcare system as a whole.
Art, an inherently personal form of expression, has become dangerously susceptible to exploitation in attempting to quantify meaning, value and emotions.
Moreover, AI models are typically trained without an understanding of cultural nuances. This can lead to misclassification of these images and harmful outcomes. For example, generative models could misinterpret images of Black individuals or communities – whether from art, social media posts or even personal cloud-stored photos – and create negative or stereotypical representations.
What once seemed like a transformative way to connect with communities is now entangled in risks of exploitation and misuse.
AI could complement art’s role in advancing health equity, but only if art is protected as any other form of data.
Art must remain rooted in human expression, unrestricted by models that fail to represent, compensate or acknowledge diverse experiences.
While a formula may claim to rival a thousand pictures, it will never surpass the value of a thousand lives. Health equity can only thrive by placing art and its profound connection to humanity at its core.
Community-based governance models, such as the Engagement, Governance, Access and Protections (EGAP) framework, might offer a path forward for responsibly integrating AI into artistic and health-related contexts. Government bodies including Innovation, Science and Economic Development Canada must prioritize consultations with artists and affected communities to design new safeguards and ensure technology benefits equity-seeking communities. With forethought and meaningful community engagement, we can chart a path that ensures art continues to support health equity.