ChatGPT in Chinese – is that possible?

ChatGPT in Chinese – is that possible?

Berlin, February 22, 2023

The release of ChatGPT, an artificial intelligence (AI) bot that answers questions in a human-like manner, in late November 2022 sparked great excitement around the world. As of January, the service had 100 million users and is already being used in a professional context: news site Buzzfeed announced in an internal memo that it would start integrating it into its processes, and it was even used by a Colombian judge in court Decision.
Microsoft has partnered with OpenAI, the company behind ChatGPT, and says it will integrate the technology into the next version of its Bing search engine. US competitors have struggled to respond. Today, February 8, Google will release its “latest and most powerful language models to complement search,” according to CEO Sundar Pichai. This follows last month’s declaration of a “Code Red” that brought all hands on deck – including founders Larry Page and Sergey Brin – to create a competitor for ChatGPT.
And other companies have suddenly released AI chatbots: Quora recently launched Poe, Anthropic has joined the fray, and You.com has had a chatbot as part of its search engine for a few months.
An AI bot for the Chinese speaking world
Excitement is also evident in China, where online discussions have ranged from excitement about the technologies to concerns about China’s ability to keep up with recent developments by US firms.
The Chinese company that stands the best chance of replicating ChatGPT’s influence is Baidu, which has been working on AI for years. The company has now confirmed that it will release its AI chatbot called Ernie Bot (or wénxīn yīyán 文心一言 in Chinese) in March. (Ernie is an acronym for Enhanced Representation through Knowledge Integration.)
But that’s all we know so far, so we can only speculate as to what Ernie Bot will be like. CE CEO Robin Li (李彦宏 Lǐ Yànhóng) exuded confidence at Baidu’s recent developer conference, but also warned that AI’s widespread implementation will take time and hinted that there are still problems in commercializing the technology give. This is true of all current AI chatbots: Google has failed to monetize its “assistant” and has reportedly cut resources from its department. Likewise with Amazon’s Alexa. And despite all the excitement surrounding ChatGPT, it doesn’t make any money.
Jeff Ding, an assistant professor at George Washington University and founder of the ChinAI substack, sees Ernie Bot as the anticipated next step in Baidu’s ongoing development of Large Language Models (LLMs). He references Ernie Bot’s predecessor, the Ernie Titan 3.0 model, which was trained on massive amounts of Chinese-language text scraped mostly from the Chinese internet.
Baidu’s Ernie line of LLMs was first introduced in 2019, building on Baidu’s previous forays into LLMS, and has now expanded to include a range of models intended for a variety of applications.
Baidu’s challenges
Ding emphasizes that unlike ChatGPT, which is aimed at an English-speaking marketplace, Chinese LLMs, including the Ernie class, Huawei’s Pangu, and Inspur’s Yuan 1.0, are optimized for the Chinese-speaking market. So it is likely that Ernie Bot will also be trained mainly with Chinese materials and will not be a direct competitor of ChatGPT or the products of other US companies.
Therein lies a disadvantage. “There’s just less quality text to train with,” says Ding. “It’s something that Chinese researchers working in this field have mentioned, and it’s also what differentiates Chinese labs from their Western counterparts.” In fact, he says, “If you think about the really high-quality academic literature out there , much of it is written in English.”

Hits: 1