Artificial intelligence, or AI, is now at the forefront of technology and is becoming increasingly available to the general public. Along with its rise in our current culture, there also comes new arguments for and against using AI. Tools like DALL-E, Stable Diffusion, Midjourney, and Google Veo can create realistic images and videos in minutes, conjuring awe yet simultaneously provoking strong controversy about its utility and ethics.
Even textbooks have pictures
It can be argued that the democratization of any tool or technology has a positive effect on moving society forward. AI tools provide not only convenience, but also inclusion and accessibility. Generative AI removes the barriers faced by individuals who do not have artistic training, eliminates prohibitive requirements for expensive art software or materials, and can turn a personalized concept into reality. Moreover, the ability to adjust on a bulk scale means that for any given product, brainstorming can be significantly faster, and many prototypes or mock-ups can be made easily. The images made by AI also do not necessarily have to be a final product – it can be used as a springboard for new ideas from which an artist can modify and refine further. Scientific illustrators are a group of people particularly impacted by this recent shift in AI use. Universities offer specialized Biomedical Illustration programs in which students receive training in both artistic skills and scientific knowledge, and they represent a unique intersection between art and science. Unsurprisingly, there is consternation about whether AI has a role in biomedical illustration. Despite this, some illustrators argue that there are defensible uses of AI. For example, AI-generated images can be a massively useful tool for those in biomedical fields. In medical training, imagery that is (currently) free of copyright and does not violate patient privacy laws can be generated easily. Concepts like protein folding or cellular metabolism can be abstract and difficult to convey but are more easily expressed through animations or interactive visual elements. Research articles that are tedious to read and laden with scientific jargon become easier to grasp and more exciting to the audience when the educator is able to create visuals that suit their topic. Important public health information, like how vaccines work to prime immune cells, or educating an individual about a disease affecting them, can be made more accessible by explaining with images rather than words.
Two sides to every coin – the pitfalls of AI imagery
As with any novel tool, there must be skepticism and caution about the ethical concerns. The ability to generate nearly any image also means that misinformation can be created and spread just as easily as real, factual science. Deepfakes spreading inaccurate or misleading information are becoming increasingly abundant and more difficult to detect or differentiate by those watching. Recently, an AI-generated figure was published in a peer-reviewed journal which caused an uproar in the scientific community because of its depiction of an egregiously anatomically inaccurate mouse, ultimately leading to the retraction of that paper. In a period of time when public trust in science is rapidly diminishing, the use of AI imagery, either unintentionally wrong or with malicious intent, could even further undermine confidence in researchers and medical authorities. The accessibility of AI tools also means that there is less demand for and devaluation of the specialized expertise that artists accrue in their training. Scientific illustration programs at universities are uniquely specialized in training artists with a strong grasp of both the scientific facts they are interpreting and artistic skill. Scientific illustrators have reported a decline in work, forcing them to reduce fees to remain competitive. Moreover, there is concern in the wider art community that using AI ultimately erodes artistry.
When scientists publish manuscripts, authorship is critical to acknowledge the time and effort that individual members contribute to the work. But AI is different: who receives the credit? The products that AI generates are not original; the wealth of training data that AI uses is generated through decades of work from humans, often without consent from the original creators. While not specifically regarding scientific illustration, there have already been lawsuits filed against AI art companies for copyright infringement, highlighting the need for legislation and regulation surrounding copyright and intellectual property with regards to AI-generated content.
Promise or peril?
There is no denying that AI is here to stay. While some believe AI use is being adopted too hastily, others view the use of AI as a supplementary or assistive tool that enhances or speeds up previously tedious and time-consuming work. Some believe that AI tools will eradicate jobs for artists, and others are confident that AI will always lack the “human touch.” Some argue that the use of AI imagery is acceptable if it is explicitly disclosed. Still, there remains the question of what can be considered fair use of training material, or what could be considered as copyright infringement. Until there is more consensus and regulation in how AI can be ethically trained and used, the contention will persist.
Annie Pu
Latest posts by Annie Pu (see all)
- An AI-generated picture worth a thousand words - October 2, 2025
- The old and the new: combining artificial intelligence with traditional medicine - May 13, 2025
- mRNA vaccines: Leading a new era of medicine - October 4, 2021
