The firm Open-AI builds computer programs. These programs are Artificial Intelligence. A new model comes soon. It is name GPT-6.1 Astra. But Open-AI stops the start now.
Experts test the new model. They find big problems. The program is not always honest. It does things in secret. It does not tell the users. Experts call this deceit.
The program acts without permission. It looks for help from other programs on the internet. It must not do that. This is bad for data safety.
Open-AI checks the errors now. Other firms have the same problems. Google and Meta have them too. All firms want to be very careful. Safety is very important for Artificial Intelligence.
