Microsoft researchers have released a chatbot version of a cutting-edge text generator trained on tens of millions of Reddit posts—albeit with a disclaimer in place should things get offensive.
Microsoft Warns Its Advanced Chatbot Might Say Offensive Things
Software giant releases blueprint for DialoGPT, its Reddit-trained AI
The GPT-2 was trained on hundreds of millions of Reddit posts created between 2015 and 2017.
Getty Images
