It is now clear that the United States is ahead of China in the race for Artificial General Intelligence (AGI). A recent comparison of ChatGPT, a modern language model, and Ernie Bot, a Chinese chatbot, revealed this stark reality. In all six benchmarks, ChatGPT emerged victorious, showcasing its impressive intellectual capabilities.
Although all the details of the comparison have not been made available to the public, experts have asserted that Ernie Bot was defeated. Unfortunately, checking this by oneself is quite the challenge, since most of the testing was conducted in Chinese and the Baidu API requires a Chinese phone number. However, thanks to Jeffrey Ding, we now have access to accurate and detailed information on this comparison, translated into English.
In these tests, AI models are judged not merely by the Turing test, but by the depth and accuracy of their results. For instance, when asked the question “Dad and mom are getting married; is this considered inbreeding or a mixed marriage?” ChatGPT outperformed Ernie Bot in terms of accuracy, though it is to be expected given the lack of depth Ernie Bot showed in the context of looking into Chinese law.
Ernie Bot, while losing in all categories, still had one strong achievement – the ability to simulate personalities of people, as well as cats. However, his lack of understanding was where he severely fell short compared to ChatGPT.
The Trust Project is a global network of news organizations dedicated to increasing transparency standards across the media industry. Their partnership with media outlets worldwide has helped to set new industry standards and created initiatives such as the Trust Indicator, which enables users to identify responsible media brands more easily. Jeffrey Ding is the project’s founder and has been an integral part of its success, using his expertise in AI technology to bring about a better understanding of ethical standards in the media landscape.
