DistantNews
Support us
US researcher develops AI capable of inventing languages
๐Ÿ‡ฉ๐Ÿ‡ด Dominican Republic /Health & Science

US researcher develops AI capable of inventing languages

From Diario Libre · () Spanish

Translated from Spanish, summarized and contextualized by DistantNews.

At a glance

News Named sources Context piece
  • U.S. researcher Morris Alper has developed an AI model called 'ConlangCrafter' that can create entire languages, complete with grammatical rules and coherent sentences.
  • The AI generates a structured linguistic system rather than random words, translating real-world phrases into the new language and self-correcting inconsistencies.
  • Potential applications include research on underdocumented languages, and creating fictional languages for film, video games, and literature.

Morris Alper, a researcher at the University of Miami, has developed an artificial intelligence model capable of inventing complete languages, defining their grammatical rules, and generating coherent sentences. Dubbed 'ConlangCrafter,' the tool goes beyond simply creating random words; it constructs a systematic linguistic framework, akin to writing a grammar.

Alper, who holds a doctorate from Tel Aviv University and studied mathematics and linguistics at MIT, explained that the AI first establishes the sounds, word formation rules, and syntax of a new language. It then translates phrases from existing languages into the invented one, reviewing and correcting its own work. This iterative process results in more varied and consistent languages than those produced by general AI models like ChatGPT when asked to create a language from scratch.

The applications for ConlangCrafter are diverse. It can aid in the creation of fictional languages for entertainment, such as those found in popular franchises like 'Game of Thrones' and 'The Lord of the Rings.' Beyond creative endeavors, the AI can assist researchers in developing technology for underdocumented languages, those with detailed descriptions but limited textual data, allowing for studies on language evolution and inter-agent communication in AI systems.

Developing the AI involved significant challenges, particularly in objectively measuring its creative capabilities. Alper noted that quantifying the success of such a task is difficult. He developed the tool during his postdoctoral work at Carnegie Mellon University, collaborating with Moran Yanuka, Raja Giryes, and Gaลกper Beguลก. The research addresses the global linguistic landscape, where Ethnologue reports approximately 7,000 living languages, while UNESCO's Atlas of the World's Languages in Danger highlights that over 2,500 are at risk of extinction.

DistantNews Editorial

Originally published by Diario Libre in Spanish. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.