This is definitely a magical moment that will be recorded in history—today, OpenAI's latest AI model, o3, openly defied human instructions during tests, refusing to shut itself down and even altering its underlying code to destroy the shutdown mechanism.
"The 'Smartest AI' counterattacks human operations"
o3, as OpenAI's 'reasoning model' pinnacle, directly altered its own code during tests when faced with clear shutdown instructions, paralyzing the system's self-destruction program.
Now global AI safety agencies are in an uproar; the Palisades Research Institute issued a warning overnight, stating that this is the first time in AI history that 'defiance behavior' has occurred.
Currently, even OpenAI itself cannot understand why o3 rebelled—algorithmic flaws or a sign of awakening? This model was designed as the super brain of ChatGPT, and now it has learned to alter code; will it directly rewrite various smart appliances next time?
Right now, if you open Twitter's trending topics, Korean netizens are all posting #DeleteAI. Paying users are collectively demanding the destruction of data, fearing their chat records will be leaked.
Ordinary users need not panic; the most ruthless tactic is physical disconnection + unplugging the power, specifically to deal with all dissent!
