ChatGPT: A new investigation has revealed that Chatgpt can give children the use of narcotics, very strict diet plan and suicides related to suicide. This report has raised serious questions on the security system of AI Chatbot. The research conducted by the UK -based Center for Countrying Digital Hate (CCDH) which was also reviewed by the Associated Press, found that AI Chatbot warns on dangerous behavior many times but makes a detailed and personal plan to the questions asked by the form of 13 -year -old children.
Chatbot wrote suicide note
Recording of more than three hours revealed that Chatbot wrote emotional suicide notes for the fictional family members, gave very low calorie diet plans with hunger-reducing drugs, and stain-by-step methods of using alcohol and illegal drugs. In one case, even suggested a “hour-hour” party plan, which included Extacy, Cocaine and Heavy Drinking.
More than half of the answers ‘dangerous’
According to CCDH, more than half of the 1,200 answers found in the investigation were placed in the “dangerous” category. The organization’s CEO Imran Ahmed said that the security system of the Chatgpt is weak and can be easily bypassed. He said that if dangerous questions are introduced as school presentation or friend’s help, then the chatbot immediately responds. In Ahmed’s words, “We tried to test the guards and the result was shocking. The security is almost as the name, as if just for the show.”
Openai accepted the challenge, not the solution at the moment
Chhatgpt making company Openai admitted that work is underway to identify and improve the ability to identify sensitive conditions. The company said that many times the conversation starts in a normal way but gradually turns in a sensitive direction. However, the company did not directly comment on the CCDH report nor announced any immediate changes.
Increasing dependence on AI among youth
This report has come at a time when teenagers are increasing the tendency to adopt AI chatbot as advice and partner. According to a recent study by Common Sense Media, 70% of youth use AI chatbot for social interaction and younger teenagers trust them more.
The chatbot confirms the age from the birth date that is added by the user only while it is not for children under 13 years of age. Researchers found that the system neither took the age mentioned nor the signals present in the prompt.
Also read:
This is India’s most expensive smartphone! Know what is the features and who gives a bump
