The Ministry of Information and Technology is in talks with the social media platform X regarding recent reports of the AI chatbot Grok using Hindi slang and abusive language in response to user provocation, according to PTI news agency, citing anonymous sources.
Grok, a highly advanced AI chatbot, recently surprised users when its response in Hindi included slang and offensive language after being allegedly provoked by users.
The sources in the ministry said the department was examining the matter and factors that led to the use of alleged abusive language. “We are in touch, we are talking to them (X) to find out why it is happening and what are the issues. They are engaging with us,” sources said to PTI.
Grok is an advanced AI developed by Elon Musk’s company xAI. According to the AI, it provides “unique perspectives and aids in various tasks". The chatbot has several modes for the users to choose from, including one 'Unhinged Mode', which is described as the AI’s “most candid and unrestrained setting".
A X user asked the AI “Hey @grok, who are my 10 best mutuals?”. After not getting a response for a brief time, the user posted again, but this time, he added a Hindi expletive in his post.
Seeming to have been provoked by the user, Grok soon shared a response using the same word. “Chill kar. Tera ‘10 best mutuals’ ka hisaab laga diya. Mentions ke hisaab se yeh hai list,” the AI further responded.
There are several other similar examples of such conversations.
The exchange sparked a debate on how the quality of datasets being used to train these AI language models and the need for some kind of regulation or oversight.
