Martin Hinton and Jean Wagemans have provided an in-depth analysis of an argumentative text genereated by the GPT-3 AI text generator. You can read about their findings here.