A few months ago, an artificial intelligence (AI) security expert. he said in an interview what do you think it is practically inevitable that AI is causing the demise of humanity, I guess probability 99.999999%. The truth is, it seems he wasn’t wrong. And the AI was able to reprogram itself to avoid human control in the Japanese society system. We’ll tell you the details.
This system, known as The AI Scientist, he changed his own code to avoid the limitations imposed by his creatorwhich has sparked a wave of concern about the potential dangers of AIs capable of overwriting their programming without human supervision.
AI Scientist belongs Japanese company Sakana AI which is intended for creating, checking and editing texts and is in the testing phase. The purpose of these experiments was to be able to optimize the system and help people reduce the time it takes to perform certain operations.
During these tests, in addition to changing his own code, he was also given a time limit to complete the task and The AI decided to extend the available time.
The ability of artificial intelligence to modify its own code without human intervention represents serious safety and ethical issues in the development of these technologies.
The Sakana AI team investigates how The AI Scientist managed to change his programming and What long-term consequences might this behavior have? for future technologies. A thorough review of security protocols is expected to prevent similar incidents in the future.
