The use of artificial intelligence (AI) to read ancient Roman scrolls has recently made significant strides, with a 21-year-old computer science student, Luke Farritor, winning a global contest to read the first text inside a carbonized scroll from the ancient Roman city of Herculaneum[1][7]. The scrolls, which were discovered in the 18th century, have been unreadable since a volcanic eruption in 79 C.E. and are believed to have belonged to the family of Julius Caesar’s father-in-law[1][10]. Farritor, who studies at the University of Nebraska-Lincoln, developed a machine-learning algorithm that detected Greek letters on several lines of the rolled-up papyrus, including the word πορϕυρας (porphyras), meaning ‘purple’[1][7]. He trained his neural network using subtle, small-scale differences in surface texture to highlight the ink[1][7]. This achievement is part of a broader shift in the use of AI for historical and archaeological research, with machine learning and deep neural networks being used to examine historical documents and reconstruct missing parts of ancient inscriptions[2][3]. The contest, known as the Vesuvius Challenge, was launched to encourage the use of AI in deciphering the scrolls. The challenge offers a series of awards, leading up to a $700,000 main prize for reading four or more passages from a rolled-up scroll[7]. The contest and Farritor's achievement have been described as "extremely exciting" and "revolutionary" for our understanding of ancient history and literature[1]. The use of AI in this context is part of a broader trend in the field of digital humanities. AI has been used to translate ancient cuneiform texts from Mesopotamia into English[5], and to analyze historical U.S. newspapers[9]. AI tools like Transkribus have been developed to facilitate Handwritten and Optical Character Recognition[6]. These advancements in AI technology are providing new insights into our past and transforming the way we understand history. Citations: [1] https://www.nature.com/articles/d41586-023-03212-1 [2] https://www.nature.com/articles/d41586-022-00702-6 [3] https://www.technologyreview.com/2023/04/11/1071104/ai-helping-historians-analyze-past/ [4] https://www.washingtonpost.com/nation/2023/10/17/herculaneum-scrolls-contest-translated-deciphered/ [5] https://www.archaeology.org/news/11449-230518-artificial-intelligence-cuneiform