Google has introduced three innovative AI-powered experiments designed to offer more personalized and effective language learning experiences. Revealed on Tuesday, these early-stage tools harness Google’s multimodal large language model known as Gemini and could position the company as a competitor to established platforms such as Duolingo.
The first experiment, named “Tiny Lesson,” specifically targets the common frustration learners face when confronted with an immediate need for a certain phrase they haven’t yet mastered. With Tiny Lesson, users describe a situation they might encounter—such as losing their passport—and immediately receive targeted grammar and vocabulary tailored precisely to that scenario. Additionally, the tool suggests culturally appropriate responses, like “I don’t know where I lost it” or “I want to report it to the police.”
Google’s second experiment, “Slang Hang,” aims to help learners transition from formal textbook speech to more natural, colloquial expressions. It simulates authentic conversations between native speakers, which users can follow and analyze one line at a time. Through realistic scenarios—such as a customer interacting with a street vendor or two friends reuniting on a subway—learners can familiarize themselves with local slang, idioms, and natural conversational flow. Users also have the option to hover over unfamiliar slang to see descriptions and contextual usage. Google acknowledges that “Slang Hang” occasionally mixes in incorrect usage or even invents slang terms inadvertently, cautioning users to verify phrases against trusted sources.
The third experiment, “Word Cam,” leverages visual recognition technology to enhance vocabulary learning by allowing users to photograph their environment. Gemini then recognizes objects in the image and labels them in the user’s target language, even suggesting additional descriptive terms. Google notes that sometimes basic knowledge of a particular word, such as “window,” might overlook essential associated terms like “blinds,” highlighting how visual context helps users discover vocabulary gaps.
These AI experiments reflect Google’s interest in exploring how advanced technologies can support dynamic and tailored individual learning experiences. Initially accessible through Google Labs, the language tools currently support a variety of languages, including Arabic, Chinese (Mainland, Hong Kong, Taiwan), English (U.S., U.K., Australia), French (Canada, France), German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Portuguese (Brazil, Portugal), Russian, Spanish (Latin America, Spain), and Turkish.
