China’s AI Bot Ernie Raises Concerns over Censorship and Selective Responses

Author: Shannon Thaler

Published: May 19th, 2023

Click to read more →

China's AI chatbot, Ernie Bot, developed by Beijing-based tech firm Baidu, has been scrutinized for banning users who ask sensitive questions about President Xi Jinping or Winnie the Pooh. In addition, Ernie Bot has been found to avoid answering questions about the origin of COVID-19. Launched in March, this AI chatbot claims to be more suited for tasks like question answering and dialogue generation when compared to OpenAI's ChatGPT. The bot's selective responses have sparked discussions about regulating AI-powered bots. When faced with questions about the origin of COVID-19, Ernie Bot exhibits evasive behavior by providing vague responses or attempting to change the topic. By avoiding any mention of China's potential role in the pandemic's origin, the chatbot raises further questions about transparency and accountability in the AI landscape. In a self-comparison with OpenAI's ChatGPT, Ernie Bot asserts its specialization in specific tasks such as question answering and dialogue generation. In contrast, ChatGPT possesses a more general ability to understand and generate natural language. This distinction highlights the diverse capabilities of AI-powered chatbots and their specific areas of expertise. The incident underscores the potential for AI technology misuse and manipulation, demanding a robust framework to ensure transparency and accountability. As AI advances, responsible regulation becomes imperative to safeguard against biased or censorial practices.