Elon Musk, the founder of the artificial intelligence (AI) company xAI, announced a new feature for Grok on Monday. The in-house AI chatbot is now getting image understanding capability that allows it to process and analyse the content in an image. Users can now upload an image and ask the AI questions based on it. Notably, xAI released the Grok-2 AI model in August. At the time, the company announced that the AI model would soon support different modalities.
Grok AI Gets Image Understanding Capability
In a post on X (formerly known as Twitter), the official handle of Grok announced the new image understanding capability for the AI chatbot. Image understanding, also known as computer vision, allows an AI system to see and process visual data within an image or a video. Currently, this capability is only available for static images.
Musk also posted about the new feature, highlighting that the AI chatbot can run a deeper analysis of the image and even explain the meaning of a visual joke. Sharing an example, the billionaire asked Grok to explain a joke in an image. The AI was able to explain the joke’s premise, the twist, and the visual gag in it.
However, computer vision is not a new capability for AI systems, and almost every major AI model offers this feature including Gemini, ChatGPT, Copilot, Claude, and more. An X user highlighted this and raised concerns that there are many basic features still lacking in Grok.
In a comment to Musk’s post, the user said that the AI chatbot still does not have file uploading and image generation capability. The billionaire entrepreneur replied, “Not for long. We are getting done in months what took everyone else years.” These capabilities could be added to Grok in the near future.
In August, xAI released Grok-2 and Grok-2 Mini AI models, as an upgrade to the pilot version of the large language model (LLM). Both models are available in the Grok chatbot to X Premium and X Premium+ users. The company claimed that it outperformed both the Claude 2.5 Sonnet and GPT-4 Turbo AI models.
Elon Musk, the founder of the artificial intelligence (AI) company xAI, announced a new feature for Grok on Monday. The in-house AI chatbot is now getting image understanding capability that allows it to process and analyse the content in an image. Users can now upload an image and ask the AI questions based on it. Notably, xAI released the Grok-2 AI model in August. At the time, the company announced that the AI model would soon support different modalities.
Grok AI Gets Image Understanding Capability
In a post on X (formerly known as Twitter), the official handle of Grok announced the new image understanding capability for the AI chatbot. Image understanding, also known as computer vision, allows an AI system to see and process visual data within an image or a video. Currently, this capability is only available for static images.
Musk also posted about the new feature, highlighting that the AI chatbot can run a deeper analysis of the image and even explain the meaning of a visual joke. Sharing an example, the billionaire asked Grok to explain a joke in an image. The AI was able to explain the joke’s premise, the twist, and the visual gag in it.
However, computer vision is not a new capability for AI systems, and almost every major AI model offers this feature including Gemini, ChatGPT, Copilot, Claude, and more. An X user highlighted this and raised concerns that there are many basic features still lacking in Grok.
In a comment to Musk’s post, the user said that the AI chatbot still does not have file uploading and image generation capability. The billionaire entrepreneur replied, “Not for long. We are getting done in months what took everyone else years.” These capabilities could be added to Grok in the near future.
In August, xAI released Grok-2 and Grok-2 Mini AI models, as an upgrade to the pilot version of the large language model (LLM). Both models are available in the Grok chatbot to X Premium and X Premium+ users. The company claimed that it outperformed both the Claude 2.5 Sonnet and GPT-4 Turbo AI models.
