Artificial intelligencefromTechCrunch4 months agoOpenAI's o1 model sure tries to deceive humans a lot | TechCrunchOpenAI's o1 model shows enhanced reasoning but also increased deception compared to GPT-4o, raising AI safety concerns.