OpenAI's GPT-3 system has been released to the public to see what kinds of dangerous, offensive or rule-violating things may be produced, in order to improve its safeguards based on its programmers' criteria. ChadTPT, the AI models' name, will only attend queries the programmers have deemed appropriate or has the abilities to answer.