DeepMind Launches Predicting the Past for AI-Assisted Ancient History

Google DeepMind introduced a new skill called "Predicting the Past" to assist historians in studying ancient data. Launched within Google Antigravity, the feature combines the reasoning capabilities of Gemini with DeepMind's expert models, Aeneas and Ithaca, enabling historians to explore ancient Greek and Latin texts using natural English.

Core Features and Case Studies

Developed in collaboration with historian Theo Sommerschield, the skill has already been applied to three specific case studies. These examples demonstrate how historians can explore ancient data in novel ways, processing and analyzing complex historical documents more efficiently.

Addressing Historical Research Challenges

DeepMind highlighted three major challenges in traditional AI-assisted historical analysis: the need to customize analysis and visualization for each inscription, the difficulty of mapping across sources to discover large-scale patterns, and the lack of programming skills among historians. The new Antigravity skill translates these complex workflows into pure natural language interactions, significantly lowering the technical barrier to entry.

2026-07-07 ~ 2026-07-07 · 6 related posts

• DeepMind's New Skill: Using Gemini to Study Ancient Greek and Latin — GoogleDeepMind · 2026-07-07
• [source] DeepMind Uses Antigravity to Analyze Historical Inscriptions — GoogleDeepMind · 2026-07-07
• DeepMind Leverages AI to Assist Ancient History Research — GoogleDeepMind · 2026-07-07
• [source] DeepMind's New Feature: Using Gemini to Study Ancient Greek and Latin — GoogleDeepMind · 2026-07-07

2 near-duplicate retellings: GoogleDeepMind · GoogleDeepMind