Ciudadanía Italiana

Study Re-evaluates Five Major Extinctions: 3 Were More Gradual

AI analysis of ~4 million fossils suggests three of the five major extinctions were more gradual, assessing over 110k species across 540 million years.

Study Re-evaluates Five Major Extinctions: 3 Were More Gradual
Foto: Daniel Ventura (CC BY-SA 3.0)

A study re-evaluating the pace of the great mass extinctions concludes that three of the five largest biodiversity losses may have occurred in a more gradual way, with slow declines rather than sudden pulses. The analysis encompasses nearly four million marine fossils and estimates origination and extinction times for over 110,000 species across 540 million years, according to a report by ANSA released on October 22, 2026.

Key conclusions — mass extinctions and the pace of events

  • Three of the five major mass extinctions analyzed show signs of prolonged declines in species diversity, in contrast to the classic image of instantaneous cataclysmic events.
  • The authors state that this revision of the historical narrative could alter interpretations of causes and dynamics of past biological crises, highlighting slower loss patterns that may have been masked by gaps in the fossil record.
  • According to ANSA, the results were presented at the Geological Society of America meeting in Denver and disseminated on the Nature journal website, where the authors discuss implications for paleobiology and conservation.

Methodology and data used — marine fossils, AI and OneStrata

  • The database includes about 3.9 million marine fossils collected from tens of thousands of sites around the world, spanning roughly 540 million years.
  • The researchers applied a new artificial intelligence model called OneStrata to infer the origination and extinction dates of species from an incomplete fossil record.
  • With OneStrata, the team estimated times for more than 110,000 species, enabling temporally resolved reconstructions of biodiversity at an unprecedented scale.
  • The approach combines statistical modeling of sparse records with machine learning techniques to reduce biases in the fossil record, according to the authors cited by ANSA.

Team and dissemination — who led the study

  • The research was conducted by a team led by researchers from Nanjing University, with participation of Chinese scientists and international collaborators, as reported by ANSA.
  • The results were publicly presented at the annual conference of the Geological Society of America (GSA) in Denver, and also made available on the journal Nature website, where science communication reaches a broad audience.
  • The combination of a large database and a new computational model made it possible to scale analyses that were previously limited to regional studies or smaller taxonomic groups.

Relevance for Brazilian readers and Italian descendants — biodiversity and science communication

  • The revision of the pace of extinctions has a direct impact on the global understanding of the history of life, information of interest to paleontologists, conservationists, and science communicators in Brazil.
  • For Brazilian audiences and the Italian-descendant community, the results illustrate how long-term data and new methodologies can reshape established scientific narratives.
  • The use of large fossil datasets and tools like OneStrata demonstrates pathways for future projects in mapping and protecting biodiversity, including in national contexts.
  • Readers interested in science-society relations can find related coverage in sections like News from Italy and reflections on life in Italy in Life in Italy. For citizenship topics and connections with Italy, see Italian Citizenship.

Limitations and next steps

  • The authors acknowledge that inferences from the fossil record depend on available sampling and model assumptions; geographic or taxonomic gaps can influence conclusions.
  • Future studies should validate OneStrata's conclusions with other approaches and expand the coverage of terrestrial fossils and underrepresented groups to test the generality of the gradual extinction pattern.
  • The reinterpretation of the pace of extinctions may spur new research on causes (climatic, tectonic, oceanic) and on how slow signals of decline relate to environmental factors.

Conclusion The AI-driven analysis of millions of fossils sheds light on a more complex picture of the big extinctions: in three of the five cases studied, the loss of diversity appears to have been an extended process, not merely an abrupt cut. The work, presented at the GSA and reported by Nature, opens avenues for historical revisions and for long-term conservation discussions.

Source: ANSA

Leer también