Welcome to Technology Thursdays on the Deep dAIve podcast!

This is our weekly slot where human curiosity meets AI execution to navigate the engineering and digital frontier. We’re independent science enthusiasts learning out loud, openly leveraging AI models to help us interpret data and write our scripts; making today's study incredibly close to home.

In this episode, we unpack a fascinating human-computer interaction study from the National University of Singapore. We all know that social chatbots make mistakes and hallucinate inaccurate information. But when they mess up, how they fix those mistakes completely dictates whether humans will ever trust them again. The researchers conducted a controlled experiment comparing three error-correction methods: an external webpage retraction, an external "expert" chatbot stepping in, or the original chatbot catching and correcting its own mistake. The data reveals a massive psychological win for AI self-correction, showing it's the only strategy that corrects misinformation without absolutely destroying the bot's long-term credibility, perceived expertise, and social connection with the user.

Read the Original Research Paper:

• Article Title: Correct Yourself, Keep My Trust: How Self-Correction and Social Connection Shape Credibility in Social Chatbots

• Authors: Biswadeep Sen and Yi-Chieh Lee (National University of Singapore)