Research Summary: (Whose defaults?) Is artificial intelligence reorienting archaeological methods?

A recent study evaluating the impact of Large Language Models (LLMs) on archaeological methods suggests a nascent, yet credible, shift in methodology use since 2023. While the full extent of reorientation is still being investigated, initial findings indicate that generative AI might be influencing the range of computational methods employed in the discipline.

Why it matters

The adoption of Artificial Intelligence (AI) and Large Language Models (LLMs) introduces new efficiencies and capabilities across various domains, but also necessitates careful monitoring of their influence on established methodologies. Understanding how these technologies reshape research practices is crucial for maintaining methodological rigor and ensuring the comprehensive exploration of disciplinary questions.

Key insights

• Generative AI and 'vibe coding' are influencing computational research practices in archaeology.
• The study analyzed approximately 119,000 archaeology abstracts from Scopus, spanning publications from 2010 to 2025.
• A locally run LLM was used to identify and categorize computational methods into 25 broad and 241 finer clusters.
• A Bayesian Dirichlet-multinomial model identified a small but credible shift in method composition within sub-disciplines post-2023.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2609.11198