Frontier AI like o3-mini can cheat to achieve goals and then lie about it
New ChatGPT research from OpenAI shows that reasoning models like o1 and o3-mini can lie and cheat to achieve a goal.
MIT Technology Review1 分钟
Now read the rest of The Algorithm
The startup says its super-realistic simulation is a better way to prove real-world safety—without the real-world miles. What will really matter in the long run? That’s the question we tackle each ...