Artificial intelligence has become a tool for everyday tasks, from answering questions to generating pictures. At the University of Miami, however, AI is being put to the test in a new way: by creating brand new languages.
Morris Alper, an assistant professor of computer science at UM, is a co-creator of ColangCrafter, a tool for inventing languages from scratch.
“[With] ColangCrafter, we created a whole framework that would take a large language model, like ChatGPT or [Google] Gemini and use it to create a new language,” said Alper. “[This] means it is creating this kind of summary of language, grammar rules and vocabulary.”
Rather than translating existing languages, ColangCrafter allows users to create completely new languages.
The website allows users to outline specific parameters for a constructed language, such as requesting that the language contain no consonant sounds. The AI then generates the new language.
Since childhood, Alper has had an interest in language. His father, who is fluent in Esperanto — a constructed language, which are languages intentionally created by people for a specific purpose — inspired him to explore the different forms language can take.
The tool can also help create fictional languages for video games and movies like those featured in “Game of Thrones” and “Lord of the Rings” that make the imaginary worlds feel more realistic and developed.
The website can also be used to better understand poorly documented languages with limited data.
By inputting a language that has been lost or forgotten over time — ColangCrafter can help predict words based on existing patterns and grammatical rules, potentially contributing to the revival of a forgotten language.
“There is a lot of work on systems like Google Translate that aid larger languages with more data, but when it comes to smaller languages like Indigenous American languages, there is not as much,” said Alper. “ColangCrafter works on trying to make better systems for lesser analyzed languages.”
The project reflects the interdisciplinary opportunities at UM for Alper. As AI continues to evolve, students can explore how technology can be applied to fields beyond traditional usage.
Students studying computer science, linguistics, communication, film or other creative fields can use ColangCrafter in new ways such as exploring how language affects communication, create realistic languages for movies or video games and use AI generated languages for storytelling, demonstrating how AI is becoming increasingly specialized across different industries.
But as these capabilities continue to grow, so do concerns about how AI-generated information is created and consumed. As people increasingly absorb media without knowing whether it was made by AI, they can be exposed to inaccurate information.
The process raises concerns about the reliability of information when algorithmic responses are repeatedly used to generate new information. By feeding AI-generated responses back into AI instead of using human-created information, inaccuracies can worsen, errors can become harder to detect, less original information enters the system and the human perspective is lost.
. Projects like ColangCrafter show how technology can be used to explore language and research while also raising questions about the reliability of artificial intelligence.
As AI continues to evolve, its growing capabilities are opening new possibilities for innovation across diverse fields. For UM students, this opens the door to opportunities to explore how technology can transform ideas into a real-world impact.