How artificial intelligence can analyze human languages through speech sounds and word structure.

A research article from the MIT new office on August 30th, 2022 had a very interesting research discussion centered around Artificial Intelligence (AI) and linguistics. Researchers from MIT, Cornell University, and McGill University have recently created an artificial intelligence system that analyzes human languages through speech sounds and word structure. The researchers didn’t give the AI any prior linguistic information because they wanted to see how much the AI can learn on it’s own. The researchers put the AI to the test by implementing problems from linguistic textbooks that featured 58 total languages. The problems had sets of words and corresponding word-form changes. The AI was capable of correctly creating a set of rules to explain the word form changes for 60% of the problems. I find this extremely impressive because the AI had no prior knowledge about linguistics and was able to easily create solutions on difficult subjects. Many linguists in the past have assumed that in order to gain a grasp on linguistics, you need to of course be human because of the human element involved in linguistics. Linguists never thought that AI could understand sound patterns and sound structure because it was thought that only humans could. The researchers than gave the AI knowledge about linguistics, similar to the knowledge that humans would obtain from a linguistic course. This resulted in the AI solving problems even more effectively. This research article shows how powerful AI can be, especially in the linguistic world. I believe that in the future, artificial intelligence will play a large role in helping people understand and learn more about language.

Leave a Reply