Successors to the popular Generative Pre-trained Transformer (GPT) neural model have been tested for morality. Any surprise they all failed?
