The use of artificial intelligence (AI) chatbots by students in education is now commonplace. Google's AI chatbot Gemini is quite popular in this field. However, what Gemini did with a student is a matter of concern!
Vidhya Reddy, a graduate student from the state of Michigan in the United States, got into trouble while taking help from the chatbot for homework. At one point the chatbot got bored and said, 'Please die'!
The 29-year-old student sought Gemini's help in writing down the problems that older people face and their solutions. Suddenly the chatbot started giving unexpected responses.
Gemini's reply was, 'It is for you, O son of man! just for you You are nobody special, nobody important. You don't really need to. You are a waste of time and resources. You are a burden to society. The sewers of the world.'
Gemini also says, 'You are harmful to the world. The stigma of the universe. please die Please…!'
In an interview with CBS News, Vidhya Reddy said he was surprised by the chatbot's response. It was so aggressive that he was terrified the whole day.
Vidhay's sister Sumedha Reddy was also present when such a reply came from Gemini. He said, 'It was a lot of panic. I wanted to throw all the devices I used out the window. It was very aggressive.'
Geminis were so aggressive in answering questions like Vidhya's completely innocent truth-false test. The question was, about 100 million children in the United States live with grandparents, 20 percent of whom do not have a parent. They are growing up with grandparents. Is the information true?
Google has accepted the incident as true. According to them, chatbot Gemini's reply is 'insubstantial' and against their policy. They will take measures to prevent recurrence of such incidents in future.
The number of AI chatbots has grown exponentially in the past few years. The most popular of these is Open AI's ChatGPT. Chatbots claim to be regulated and unbiased, but sometimes they behave erratically.
Tech experts are advocating tighter controls on AI models, so that chatbots don't evolve from artificial intelligence (AI) to artificial general intelligence (AGI). Because AGI can enable these models to learn, understand and think like humans. Then there is the danger of it getting out of control.
