Even Grok’s AI Can Now “see”
There are many trends in the field of generative artificial intelligence. There are reasoning models, such as OpenAI’s o3 , that “think through” each step of a problem before arriving at an answer. There are also ” deep research ” features that allow you to collect information from the Internet and create reports for you.
But perhaps the most “futuristic” trend of all is voice mode. This is 2013’s Her Promise : a chatbot you can chat with like any other person. The chatbot doesn’t say anything different than if you were communicating via text; however, it responds in a “realistic” and “natural” voice, which can create the illusion that you are talking to a human rather than a robot.
I never found this feature particularly interesting, even from big names like ChatGPT. The technology is certainly impressive, but it is still painfully obvious to my ears that I am talking to a bot. AI companies haven’t been able to get rid of these distinctive features , but that hasn’t stopped people from forming relationships with chatbots and even falling in love with them .
What impresses me most is the “vision” component of this feature. Some chatbots can not only talk to you, but also have access to your camera to see what you see and include that information in their responses. Both ChatGPT and Gemini offer these features, and now Grok does too.
Grok can see
Grok is the latest chatbot to get this ability in voice mode. xAI developer Abby Amir announced the feature, dubbed “Grok Vision,” on Tuesday , noting that Grok Vision supports multilingual audio as well as real-time search. However, these latest features are only available to SuperGrok subscribers .
This tweet is currently unavailable. It may be downloading or has been deleted.
This feature is already available on my side. You can access it by clicking on the existing voice mode option. If you haven’t used this feature yet, you’ll need to give Grok permission to access your device’s microphone. After this, you can immediately start communicating.
However, to access Vision, you need to tap the camera icon in the bottom left corner. Here, allow Grok access to your camera. Once the feed is active, you can start asking Grock what he sees.
I don’t really want to send the live video feed directly to xAI, so I kept the phone directly on the table, so the video feed was completely black. Grok, to his credit, genuinely tried to help me solve the problem, suggesting that perhaps there was something wrong with the camera or that it was too dark. When I told him that I actually took my phone with me into outer space, he “laughed” and concluded that this must be the problem: “Huh, space, right? That black signal now makes sense – there’s no light there, and the camera probably wasn’t designed for that kind of environment. You may need a space-grade device to get the signal right.”
This is the second major update to Grok this month. Last week, xAI introduced a memory feature for the bot , which allows it to access past conversations for more relevant responses.