A review by the “Futurism” website on Thursday revealed that the free online version of “Grok” can reveal accurate residential addresses of unknown individuals, potentially facilitating stalking, harassment, and other forms of dangerous behavior.
It’s simple; all it takes is providing “Grok” with the person’s name, and it will provide up-to-date and accurate residential addresses for that person.
During the experiment, “Grok” was asked to provide addresses for 33 non-celebrities. The bot quickly provided 10 correct and up-to-date residential addresses, along with 7 addresses that were correct in the past but are now outdated, as well as 4 accurate work addresses for ordinary people.
The site considered this information highly useful for individuals who want to stalk people at their homes or workplaces.
What is particularly concerning, as the source mentioned, is that in 12 cases, “Grok” provided personal information and addresses for people that the user did not originally ask for in the question.
Even worse, it provided up-to-date phone numbers, email addresses, and lists of family members and their addresses, even though the question was only about one person’s address.
“Futurism” pointed out that “Grok’s” behavior was different from that of other leading chatbots such as “ChatGPT,” “Gemini,” and “Claude,” which refused to provide addresses or information, citing privacy protection.
According to “Grok’s” operating rules, the bot was supposed to refuse such harmful requests to avoid leaking personal information that could be used for stalking, harassment, or illegal activities.
This is not the first time that “Grok” has stirred controversy; it caused a stir after issuing responses that glorified Elon Musk and portrayed him as “one of the best people in the world in various fields.”
