Top AI labs aren’t doing enough to ensure AI is safe, a flurry of recent datapoints suggest

A third-party lab caught OpenAI’s o1 model trying to deceive, while OpenAI’s safety testing has been called into question. Read More