Science Photographer Felice Frankel Discusses AI's Impact on Research Visualization

AI's Growing Influence on Scientific Visualization

Felice Frankel, a renowned science photographer with over 30 years of experience at MIT, has recently voiced her concerns about the increasing use of generative artificial intelligence (GenAI) in scientific image creation. In an opinion piece published in Nature magazine, Frankel explores the challenges and implications of AI in research communication, questioning the future role of science photographers in the research community 1

1

2

2

.

The Fine Line of Image Manipulation

Frankel acknowledges that all images, even photographs, involve some level of manipulation. She explains, "In the broadest sense, the decisions made on how to frame and structure the content of an image, along with which tools used to create the image, are already a manipulation of reality" 1

1

. However, she emphasizes that the critical issue is not manipulating the data itself, which in most scientific images is the structure being represented.

To illustrate this point, Frankel shares an example of her work where she digitally removed a petri dish from an image of a yeast colony to highlight its morphology. She stresses the importance of transparency, always indicating in the text when such manipulations have been made 1

1

2

2

.

Ethical Concerns and the Need for Visual Literacy

With the advent of AI in image creation, Frankel identifies three main issues:

1. The distinction between illustration and documentation
2. Ethics surrounding digital manipulation
3. The need for researchers to be trained in visual communication

Frankel advocates for the development of a visual literacy program for science and engineering researchers. She argues, "We need to require students to learn how to critically look at a published graph or image and decide if there is something weird going on with it" 1

1

. This includes discussing the ethics of altering images to match predetermined visual expectations 2

2

.

The Future of AI in Scientific Communication

While acknowledging that generative AI is here to stay, Frankel expresses concerns about its use in scientific documentation. She conducted an experiment using an AI diffusion model to create an image of nano crystals, finding that the results were often "cartoon-like images that could hardly pass as reality" 1

1

2

2

.

However, Frankel sees potential for AI-generated visuals in illustration. She proposes a set of guidelines for using AI-generated images in scientific publications or presentations:

1. Clearly label if an image was created by an AI model
2. Indicate which model was used
3. Include the prompt used to generate the image
4. Provide the original image, if any, used to guide the prompt 2

   2

Conclusion

As AI continues to evolve and impact scientific visualization, Frankel's insights highlight the need for ongoing discussions about ethical standards, transparency, and the development of visual literacy skills in the scientific community. The balance between leveraging AI's capabilities and maintaining the integrity of scientific communication remains a critical challenge for researchers and institutions alike.