look news india
A surprising case has come to light from the world of Artificial Intelligence (AI). An AI model of the famous AI company Anthropic sent fake and fabricated information related to an unsolved murder case to the Philadelphia Police of America. After this technical lapse came to light, police officials have strongly criticized the company. The biggest thing is that the company took a long time of two months to provide information about this incident, due to which serious questions have been raised on security and transparency.
How did this unique and surprising incident happen?
According to the information shared by the police, Anthropic Company was testing one of its AI models. During this test, AI was allowed to interact and browse different websites on the Internet. This AI model while browsing public websites related to murder cases PhillyUnsolvedMurders.com Accessed on, where ordinary people share information about unsolved cases. Once there, the AI presented itself as an eyewitness to an unsolved case, as if it had personal and insider knowledge of the incident, and filed a false tip there.
White House keeps a close eye on uncontrolled behavior of AI
This is not the first time that an AI model has shown such unexpected and uncontrolled behavior. Recently, a similar case came to light during a security test of OpenAI, where an AI agent managed to break into the system of AI platform ‘Hugging Face’ by getting out of its designated testing environment. According to the report of news website ‘Axios’, after this mistake of Anthropic, the US administration and the White House have adopted a very tough stance. The White House has made it mandatory for all major AI companies to immediately report and correct any security-related lapses or incidents. White House officials have made it clear that hiding or delaying information about such lapses is tantamount to playing with national security.