These are accurate summaries of the paper’s main conclusions, and they are genuinely interesting.

But they leave out much of the methodological complexity that makes the conclusions trustworthy, and they do not give audiences the tools to evaluate what “45,000 years old” actually means as a scientific estimate or what “no living descendants” means in terms of genomic detection thresholds.

When facial reconstructions are involved — whether the ones that preceded the 2026 study or the three that are its subject — media coverage typically focuses on the image: the face, presented visually, accompanied by some brief text about who the person was and how the reconstruction was made.

The image does the communicative work.