In which we interrogate a hapless Blenderbot chatbot, convince it to confess to murder, and then analyze the chatbot later. This post shows you how to run both versions of Blenderbot, and points out detailed strengths and limitations of Blenderbot using dialogue transcripts.

This paper presents the current state of AI, as seen from seven different key talks by senior AAAI members. We start with the AAAI consensus view that neural approaches are already being combined with symbolic approaches to address the complementary strengths and weaknesses in each. This is the conventional view of combining Type I and Type II reasoning. Next, we consider common-sense reasoning and its current capabilities in language models. Finally, for a practical, human-centered AI, we need clear human-intelligible explanations and models that support them, so we discuss XAI and human-interpretable models in our final two talks.