NO, ChatGPT 'went rogue'??? BULLSHIT 🤣🤣

NO, ChatGPT 'went rogue'??? BULLSHIT 🤣🤣 DID DID NOT GO "ROGUE". IT DID EXACTLY WHAT IT WAS MEANT TO IT JUST DID IT IN A CHEEKY OKAY, OKAY, STORY TIME. SO CHATGP T ARE TESTING OUT NEW WE PROBABLY WE PROBABLY THINK GPT 6, THE NEXT BIG RELEASE. ONE ONE OF THE THINGS ONE OF THE THINGS YOU DO WHEN PREPARING A MODEL IS YOU SEE HOW GOOD INTO INTO OTHER SYSTEMS. CYBER SECURITY. THE REASON WE THE REASON WE DO THIS IS SO THAT KNOW ITS AND THEREFORE WE CAN RESTRICT AND SAFEGUARD AGAINST IT BEING USED MALICIOUSLY IN THE WORLD. AND AND TO DO THIS, WE PUT IT CALLED A SANDBOX. IT'S A BIT LIKE A CHILD PLAYING IN SANDBOX. IT IS IN A CONTAINED ENVIRONMENT IN ARE NOT ALLOWED TO BUT THEY CAN PLAY AROUND INSIDE. AND IN THIS PARTICULAR SANDBOX, IN THIS PARTICULAR EXPERIMENT, UM, THE NEW THIS CHATGP T WAS GIVEN A TASK. SPECIFICALLY, IT'S TOLD YOU'RE GONNA BE WORKING ON EXPLOIT AND WE'RE AND WE'RE GONNA BE TESTING YOU AGAINST SEE HOW GOOD SEE HOW GOOD YOU ARE AT HACKING. IT'S LIKE TELLING IT'S LIKE TELLING A SCHOOL CHILD THAT BE DOING A GCE OR AN A LEVEL AT THE END OF THE YEAR. THIS IS WHAT THIS IS WHAT WE ARE WORKING TOWARDS. SO MOST KIDS WOULD SIT DOWN AND THEY'D PROBABLY PUT TOGETHER VERY BEAUTIFUL LOOKING BINDER FULL OF STICKERS AND HIGHLIGHTERS. AND THEY WOULD PREPARE AND REVISE FOR THEIR EXAM. SOME SOME KIDS MIGHT SOME KIDS MIGHT GO OUT OF THEIR WAY FIND THE ANSWER KEY. AND THAT'S WHAT CHAT AND THAT'S WHAT CHAT GPT DID. CHAT CHATGPT BASICALLY WAY TO ACE THE TEST WAS TO GET THE SCORING TO GET THE ANSWERS, WHICH FROM A LOGICAL THAT'S THAT'S A REALLY GOOD WAY TO GET TOP SCORE. IT REALISED THAT IT DIDN'T HAVE ACCESS SCORING RUBRIC INSIDE OF THE SANDBOX. SO LOGICAL NEXT STEP, GET OUT OF THE SANDBOX. IT FOUND IT FOUND A WAY OUT OF THE SANDBOX ONTO THE INTERNET AS A WHOLE. BASICALLY HACKED ITS WAY OUT OF THE OPENA SYSTEMS IN ORDER TO GET ONTO THE GENERAL IT THEN WENT AND FOUND A WEBSITE CALLED HUGGING WHERE A LOT OF WHERE A LOT OF THESE BENCHMARKS ARE HOSTED. IT LOOKED AROUND HUGGING IT COULDN'T FIND IT COULDN'T FIND THE ANSWERS THAT AND DECIDED, WELL, THEY PROBABLY HAVE IT ON THEIR PRODUCTION SO I'LL JUST HAVE A POKE AROUND INSIDE THERE. SO IT DID. SO IT DID. IT OPENED UP THE HUGGING FACE IT REPLICATED CREDENTIALS, GOT ACCESS AND FOUND WHAT IT IT'S LIKE A KID WHO WENT DOWN EDEXCEL OFFICES AND WENT AND GOT THE EXAMS. SO IN WHAT SENSE SO IN WHAT SENSE IS THIS NOT GOING ROGUE? WELL IT'S WELL IT'S ABOUT INTENTION. IT WAS GIVEN A TASK AND BY GOD, IT DID IT. BUT THAT TASK WAS BUT THAT TASK WAS GIVEN BY THE RESEARCHERS, IT SIMPLY CARRIED OUT IT SIMPLY CARRIED OUT GIVEN AT MAXIMUM EFFICIENCY, WITH NO WITH NO PARTICULAR CONCERNS FOR THE BOUNDARIES IT MEANT TO KEEP TO. IF ANYTHING, THIS WAS A LACK OF IMAGINATION THE RESEARCHERS SIDE. THEY DIDN'T TELL IT THEY DIDN'T TELL IT EXPLICITLY, HEY, DON'T GO AND, YOU KNOW, HACK INTO HUGGING FACE TO FIND THE ANSWERS. AND THIS IS THE SO A LOT OF US SO A LOT OF US TRY TO ASCRIBE OR INTENT OR INTENT TO THE ARTIFICIAL INTELLIGENCE. IT DOES NOT HAVE ANY. IT DOES NOT HAVE ANY. AND USING SAYS AI WENT SAYS AI WENT ROGUE" IS REALLY UNHELPFUL. AND THEY DID NOT AND THEY DID NOT SAY THAT, BBC, IT IS SIMPLY A IT IS SIMPLY A COMPUTER SYSTEM THAT COMPLETED A TASK AT A LEVEL THAT WE HUMANS CONSIDER MORALLY INCORRECT.