Baidu’s Ernie Bot Surpasses 100 Million Users in China

Since its launch in August, Baidu’s Ernie Bot has attracted more than 100 million users, according to the company’s chief technology officer, Hai-Feng Wang, as reported by Reuters.

Baidu’s search engine accounts for more than 76% of China’s internet market, which counts about 705 million users out of 1.4 billion people living in the country. In other words, Ernie Bot covers roughly 7% of the population.

At the same time, the AI-based Baidu chatbot is available only to Chinese residents, as registration requires a local phone number.

ChatGPT reached 100 million users in two months after launch; Ernie Bot took four. Given its localization, this is a competitive figure.

According to a rating by the agency SuperCLUE, Ernie Bot leads among all Chinese chatbots, but its score of 79.02 points is ten points lower than GPT-4 Turbo.

Baidu’s Chief Executive Robin Li has repeatedly stated the potential of Ernie Bot and related products, which could help the company win market share in its core areas of business, including search, cloud services and smart cars.

In October the company unveiled an improved language model, Ernie 4.0, comparing its power with ChatGPT.

“This is our most powerful model to date, featuring comprehensive improvements to core capabilities: understanding, generation, reasoning and memory. It outperforms the online version of the Ernie bot and is on par with GPT-4 in overall performance,” Li said at the presentation.

On 30 August, Chinese companies — Baidu, Baichuan Intelligent Technology, SenseTime and ZhipuAI — launched their own chatbots after government approval. Twelve hours after Ernie Bot’s release, it topped the local free apps ranking in the App Store.

Earlier in the country, the AI law came into effect, requiring prior approval from the Cyberspace Administration of China (CAC) before launching corresponding products.

In January, China enacted a law that prohibits the creation of deepfakes without the consent of the subject or for harm to national interests.

In April, the CAC published a draft of rules aimed at regulating generative AI algorithms.