3 Questions: Visualizing Research in the Age of AI – Insights from Science Photographer Felice Frankel

In an era dominated by rapidly evolving artificial intelligence, the visual representation of scientific research faces unprecedented challenges and opportunities. Felice Frankel, a science photographer with over 30 years of experience at MIT, delves into these complexities in a recent Nature magazine opinion piece. Frankel explores the implications of generative AI (GenAI) on research communication and the evolving role of science photography.

Frankel raises critical questions about the manipulation of images, the ethical responsibilities of researchers, and the future of visual communication in science. Here are some of the key insights from the interview:

Q: Defining Acceptable Image Manipulation:

Frankel acknowledges that all images are, to some extent, manipulated. Decisions on framing, structure, and tools inherently alter reality. However, the crucial point is to avoid manipulating the underlying data, especially the structure within an image. As an example, Frankel discusses removing a petri dish from an image of a yeast colony to highlight its morphology, clarifying that the colony’s structure, the core data, remained unaltered. Transparency is key; Frankel always discloses any image alterations in the accompanying text, a practice she also addresses in her handbook, “The Visual Elements, Photography.”

Q: Ethical Communication of Research:

Frankel identifies three major concerns regarding visual representation in the age of AI: distinguishing between illustration and documentation, the ethics of digital manipulation, and the ongoing need for visual communication training for researchers. She advocates for visual literacy programs within science and engineering curricula, emphasizing the importance of critically evaluating published graphs and images for potential distortions or unethical nudges. Frankel recounts an incident where a student altered one of her images without permission, underscoring the necessity of ethical considerations in visual communication.

Q: The Future of Visual Communication with Generative AI:

Frankel recognizes that generative AI is here to stay and can be a useful tool for illustration. She demonstrated this by using a diffusion model to create an image of Moungi Bawendi’s nanocrystals, highlighting the current limitations of AI-generated images to accurately represent reality. Frankel stresses the importance of establishing clear standards for the use of AI in scientific visuals. Specifically, GenAI visuals should never be used as documentation. When AI-generated visuals are used for illustration, researchers must:

• Clearly label the image as AI-generated.
• Specify the AI model used.
• Include the prompt used to generate the image.
• Provide any source images used to inform the prompt.

By adhering to these guidelines, the scientific community can harness the power of AI while maintaining the integrity and trustworthiness of visual research communication.