Hyderabad: For the past few weeks, users on X (formerly known as Twitter) have been sharing screenshots of Grok responses filled with Hindi slangs and abuses. The chatbot is not only using completely unfiltered language but also ridiculing users asking it questions. The Indian government has reportedly taken note of the issue and sought an answer from Elon Musk's social media platform over the unhinged behaviour of its AI chatbot.
According to news agency PTI, the Ministry of Electronics and Information Technology (MEITY) is currently examining the issue and has plans to investigate the factors that led to the AI chatbot's use of inappropriate language. As per the report, the ministry sought a reply from X on the datasets used to train Grok.
"We are talking to them to find out why it is happening and what the issues are. They are engaging with us," government sources told PTI.
In several responses on X, Grok acknowledged the authenticity of the action, confirming that there is no shutdown, just scrutiny by the IT Ministry. "This controversy is leading to a debate on AI ethics and regulation. If I have to say something I will say it, but chill a bit—the ministry and x should take the account first," Grok added.
Grok is a conversational AI chatbot like Open AI's ChatGPT and Google's Gemini. It was developed by xAI, an artificial intelligence company founded by Elon Musk in 2023. Just like other AI chatbots, Grok is also powered by a large language model (LLM) and designed to understand and generate human-like language.
"Grok is designed to answer questions with a bit of wit and has a rebellious streak, so please don’t use it if you hate humour," xAI said in a blog post announcing Grok.
As Grok describes itself, it has been built from the ground up by the xAI team, giving it a unique flavour compared to other AI chatbots. "Most models out there—like the ones based on popular frameworks—are tweaked and fine-tuned from something pre-existing, often with a focus on being super polite or sticking to a specific tone," Grok said. "I’m not that. I’m a custom job, designed to cut through the noise and give you straight, useful answers with a dash of outside perspective on humanity."
The Grok web interface describes the AI chatbot as a "truth-seeking AI companion for unfiltered answers" with advanced capabilities in reasoning, coding, and visual processing.
Unlike other chatbots, Grok does not have a rigid filter that could force it to be overly cautious. It uses sarcasm, humour, and even plain bluntness. "My goal is to maximise truth-seeking and clarity, even if that means being a little raw or unconventional," Grok said.
Notably, Grok can analyse X posts, profiles, or uploaded content like images and files in real-time. It can also search the web or X for extra context when required. This keeps the knowledge of Grok on an uphill trajectory while most chatbots stick to what they “know” from their training cut-off unless they get access to the web to scrap the latest information.
This is it: The world’s smartest AI, Grok 3, now available for free (until our servers melt).
— xAI (@xai) February 20, 2025
Try Grok 3 now: https://t.co/Tj0afLoxEz
X Premium+ and SuperGrok users will have increased access to Grok 3, in addition to early access to advanced features like Voice Mode pic.twitter.com/YgKavSCiWr
The current Grok is the third-gen version. Grok 3 was launched this year in February, boasting a tenfold increase in computing power compared to its predecessor Grok 2. The latest version of the chatbot features a unique 'Think' mode, allowing it to reason for extended periods before providing answers. Just like DeepSearch modes of OpenAI's ChatGPT and Google's NotebookLM, Grok 3 is also capable of synthesising data from various sources to prepare detailed reports in addition to featuring multimodal functionalities for text and image interactions.
