The Chinese company claims to own a chatbot that is superior to ChatGPT in some aspects

Baidu, China's top search engine provider, says its latest chatbot version has surpassed ChatGPT in several key metrics.

Prime Minister: 6 major orientations to cope with 'six headwinds' hindering world economic growth...
Is there a need for a law to control artificial intelligence
[Infographic] How to detect phishing websites

The Baidu technology company logo appears on the advertising screen. Photo: Reuters

According to Reuters news agency, in a statement on June 27, Baidu said that Ernie 3.5, the latest version of Ernie model equipped with artificial intelligence (AI), has surpassed ChatGPT in comprehensive competency score and surpassed ChatGPT. GPT-4 software on some Chinese language related capabilities.

Baidu cites an experiment conducted by China Science Daily newspaper using data sets including AGIEval and C-Eval, two standards used to evaluate the performance of AI models.

Baidu's announcement comes as the global popularity of AI initiated by ChatGPT has spread to China, prompting a host of companies in the country to announce competing products.

In March, Baidu was the first major Chinese technology company to launch an AI chatbot product to compete with ChatGPT. Ernie Bot, built on Baidu's Ernie 3.0 AI model, has been tested for the past three months.

Besides Baidu, other major Chinese tech companies, including Alibaba Group and Tencent, have all revealed their respective AI models.

Baidu says the new AI engine achieves better inference and training efficiency, helping to process iterations faster and with less cost in the future. Baidu also said the new model will support external "plugins".

Plugins are additional applications that allow Baidu's AI to work in specific situations such as summarizing long texts and generating more accurate answers.

Also in March, ChatGPT rolled out plugin support.

AI chatbot Q&A software became the focus of the world when ChatGPT, developed by OpenAI (USA) launched in November 2022. Since then, the tool has attracted 100 million users, setting a record as the fastest growing application. The AI-powered engine is smart enough to pass the exams for law and medical school.

In a commentary in February, cybersecurity expert MH Homaei noted that the AI ​​software poses "a number of security risks that organizations must be aware of in order to protect sensitive data and identities." language". The effectiveness of a chatbot depends on the customer providing it with valuable data, and this can make it vulnerable to information leaks. On March 26, OpenAI revealed that some users' personal and payment information had been leaked for a short time due to a system error.