Every human invention has its good and bad sides. Teflon keeps food from sticking to surfaces and makes bags waterproof, but the PFAS it produces is now contaminating the bodies of everyone in the world. Artificial intelligence (AI) helps lighten the load in the office but is now causing thousands to be laid off and students to cheat in school. OpenAI today. published a report that their new AI model is capable of producing biological weapons.
The new model was originally developed to help scientists test the latest drugs and make new discoveries. But it could also be used to ask AI to produce biological weapons if used by irresponsible people. Recognizing this potential risk, OpenAI said its AI model for biological purposes is now being developed with several safety features.
Among the things it does is train the model to reject dangerous requests. It also has a built-in detection system that tracks activity that may be related to producing biological weapons. If such activity is detected, user accounts will be blocked.
OpenAI also said that they are testing various techniques that could be used to bypass security systems. Apart from that, security systems are also used to prevent these AI models from being stolen by certain parties.