Artificial intelligence's interpretation of the neuroanatomical aspect of Peter Paul Rubens's copy of “The Battle of Anghiari” by Leonardo da Vinci

The role of artificial intelligence (AI) in art is growing every day. In general, research in the field of AI involves creating machines that have the ability to perceive, learn, and reason (Macpherson et al., 2012).

It is worth noting the achievements of AI in art. There is much debate as to the extent to which the paintings (Chamberlain et al., 2018) and music (Moffat & Kelly, 2006) generated by AI are valuable. Research suggests that people often cannot differentiate between AI-generated and man-made art (Gangadharbatla, 2022).

In 2022, we published a description of a neuroanatomical interpretation of Peter Paul Rubens's copy of “The Battle of Anghiari” by Leonardo da Vinci (Keshelava, 2022). This time we tried to see how the same work would be interpreted by AI, when prompted to find a neuroanatomical aspect.

In 1503, Leonardo created the cartoon and preparatory sketches for the vast fresco of “The Battle of Anghiari.” In 1505, he began to paint the fresco, but the experimental technique used by the master proved unsuccessful. The cartoon was cut up and sold. Its location remains a mystery. Before the disappearance of the cartoon several artists managed to create a copy. Among these copies, the drawing made by Rubens in 1603 is considered the best (Figure 1).

OPEN IN VIEWER

Figure 1.

Peter Paul Rubens's copy of “The Battle of Anghiari” by Leonardo da Vinci.

Interpretation of the Drawing: Our goal was to see how AI would interpret this drawing and how well its perception matched our own in this particular case. We used WOMBO Dream, an AI-based algorithm that can create new images by combining text and image prompts provided by a user. The keyword we provided for the algorithm was “brain” and the reference image was Rubens's drawing (Figure 1). We have not provided any other information for the algorithm. The image generated by AI is presented in Figure 2A. If we compare this image with the image obtained as a result of our interpretation (Figure 2B), we will see a sharp similarity. AI, like us, perceived the whole drawing as a representation of the brain. The image generated by the algorithm appears to show the hemisphere, cerebellum, and brain stem. This is only one possible interpretation by the AI, and different generated images could be obtained by using different text prompts.

OPEN IN VIEWER

Figure 2.

(A) The image is generated by artificial intelligence (AI). (B) Image of the brain by our interpretation.

Neuroanatomy and High Renaissance Masterpieces: Renaissance masterpieces have attracted the attention of many researchers in a neuroanatomical context, due to the fact that people of this era were interested in the anatomy of the brain. Until the 16th century, knowledge about the brain was limited to that in the teaching of Galen (Schiller, 1997). As far as is known, his anatomical knowledge was based on animal dissection (Finger, 2000). During the Renaissance, scholars began to study the anatomy of dissected human cadavers (Paluzzi et al., 2007).

Paluzzi et al. interpreted Rafael's “Transfiguration of Jesus” and Gerard David's “Transfiguration of Christ.” In both works, they saw hidden brain images (Paluzzi et al., 2007).

In 1990, Meshberger proposed a neuroanatomical interpretation of Michelangelo's “Creation of Adam” (Meshberger, 1990). According to some researchers, the images of God on the Sistine Chapel portray a range of animal brains and echo a modern interpretation of brain evolution (Ashford & Tatem, 2022). A number of studies have suggested that Michelangelo concealed anatomic information about the human brain in his masterpieces (Savastano et al., 2022).

From the above, it can be seen that several of the paintings of the Renaissance contain hidden brain images. It is also noteworthy to find other secretly described anatomical elements in the works of this era (Keshelava, 2021). Because there is no documentary evidence of these artist's intentions, the interpretations produced by researchers always lead to differences of opinion.

We think there is a possibility that AI can play an important role in refining such interpretations. An example of this is an AI's interpretation of Ruben's copy of “The Battle of Anghiari” by Leonardo da Vinci presented here.

It is interesting to wonder how the renaissance masters would have used the generative AI tools we now have access to. Massimo Paltrinieri, an expert in strategic consulting in Digital Transformation, author of technology patents (the US and EU), and former strategic consultant for a leading US company, said in an interview: “If Leonardo lived today, he would develop a higher intelligence than humans right now, leading us directly to the technological singularity.”