ARTFEED — Contemporary Art Intelligence

AI Analysis Suggests El Greco Painted Majority of 'The Baptism of Christ'

ai-technology · 2026-04-21

Artificial intelligence has challenged long-held beliefs about El Greco's painting 'The Baptism of Christ,' installed in 1624 at the Hospital Tavera church in Toledo. Researchers from Western Reserve University in Ohio published findings in Science Advances on April 17, using a machine-learning model called Patch to analyze the work at a microscopic level. The model, trained on 25 paintings by nine student artists, scanned high-resolution 3D images to compare texture patterns across the canvas. Contrary to scholarly assumptions that El Greco's son Jorge Manuel and workshop assistants contributed significantly, Patch indicated underlying connections suggesting El Greco painted most of it himself. The analysis also examined 'Christ on the Cross' (1600–10), confirming it as the work of a single artist. While a region at the bottom shows later hands, stylistic variations may stem from El Greco's experimentation, brush choices, or aging effects. Patch operates without external training data, offering potential to attribute Renaissance workshop contributions more accurately. Lead author Andrew Van Horn noted further work is needed to identify artistic practice regimes conclusively, which could track artists' development from apprenticeship to independent work.

Key facts

  • El Greco's 'The Baptism of Christ' was installed in Hospital Tavera church in Toledo in 1624.
  • AI analysis suggests El Greco painted the majority of the work, contrary to previous attributions to his son Jorge Manuel and workshop assistants.
  • Research was led by Western Reserve University in Ohio and published in Science Advances on April 17.
  • The machine-learning model Patch uses high-resolution 3D imaging to analyze brushstroke textures.
  • Patch was trained on 25 paintings by nine student artists before being applied to El Greco's works.
  • The model confirmed 'Christ on the Cross' (1600–10) as the work of a single artist.
  • Patch does not require external training data, known as 'ground truth' in AI terminology.
  • Lead author Andrew Van Horn stated further research is needed to identify artistic practice regimes.

Entities

Artists

  • El Greco
  • Jorge Manuel
  • Andrew Van Horn
  • Mark Hamilton
  • Richard Taylor
  • David M. Ewalt
  • Jackie Flynn Mogensen
  • Claire Cameron

Institutions

  • Western Reserve University
  • Science Advances
  • Hospital Tavera
  • Purdue University
  • Massachusetts Institute of Technology
  • Computer Science & Artificial Intelligence Laboratory
  • University of Oregon
  • Scientific American

Locations

  • Toledo
  • Spain
  • Ohio
  • United States

Sources