🚨 Big news: OpenAI's GPT-4o mini debuts with "instruction hierarchy" safety technique to prevent misuse and unauthorized commands. A step forward in AI safety? 🤖🔐 #AI#OpenAI#TechNews#SafetyTech#Innovation