Testing Out Google’s Gemini Vision Capabilities

Listener Brian Clark demonstrates the capabilities of Google’s Gemini Vision AI in assisting visually impaired users. Brian starts with a live demo, asking Gemini to describe his surroundings: “I see a desk with a few items on it. There is a black mug, a microphone on a stand, a keyboard, a red cable, and two laptops,” Gemini responds. It correctly identifies objects like an Echo speaker and an old iPhone on the desk.

Brian tests Gemini’s reading ability by pointing it at his computer screen, where it reads aloud from a document: “It is the grief that binds us… those that have lived through a shared vision loss have a shared experience of grief.” He then walks down a hallway, with Gemini providing navigation assistance by describing doors and the layout of a kitchen, including the exact location of the sink and items inside the fridge, such as “mustard on the second shelf, left side.”

In a final test, Brian holds up various canned goods without rotating them, and Gemini accurately identifies each one, from Campbell’s chicken noodle soup to Bush’s baked beans. “I didn’t spin any of the cans around, and it still figured out what was what,” Brian remarks, impressed by the AI’s object recognition. Gemini Vision’s ability to provide real-time assistance for everyday tasks highlights its potential as a valuable tool for the visually impaired.