Connecting links with code: AI in morphological evolution

Evolutionary biology and paleontology are based on fossil material of morphological data, i.e., anatomical, size, symmetry, ontogenetic, etc., as the dominant empirical record of the history of life over deep time. Such physical records tie evolutionary histories to the geologic past, guiding evolutionary molecular clock calibrations, phylogenetic speculations and macroevolutionary trends such as adaptive radiations, mass extinctions, morphological stasis and disparity. Additional explanations of the history of biogeography, evolution inside niches and lineage adaptation to environmental disturbances are provided by spatial-temporal variability in the fossil record. Although central, traditional morphological studies are characterized by methodological shortcomings: manual scoring of characters, linear measurements and landmark-based geometric morphometrics which causes subjectivity, homoplasy and absence of data and topological instability in phylogenies. The present review features fields in evolutionary biology and paleontology that can be improved by AI and machine learning developments in processing morphological data, which quantitative shape analysis automates, reduces biases and scales inference, which can prove. We have discovered technical and methodological gaps and have mapped out ways of integrating these tools into strong, repeatable workflows.