JD.com Inc., the largest online retailer in China, has announced plans to introduce an artificial intelligence chatbot with similar features as OpenAI LLC’s ChatGPT.

CNBC reported the development today. JD.com’s chatbot will reportedly be known as ChatJD and is described as an “industrial version” of ChatGPT. It’s expected to place a particular emphasis on use cases in the retail and financial sectors.

According to CNBC, ChatJD could help online retailers automatically generate product summaries for their e-commerce websites. The chatbot is reportedly also capable of “helping with financial analysis.” To give ChatJD an edge over rival services, JD.com reportedly plans to draw on its expertise in e-commerce, logistics and financial services.

JD.com has over the years expanded its focus beyond e-commerce to a range of other industries. The company now has a presence in the healthcare, logistics, cloud computing and physical retail segments, among other areas. It generated revenues of $34.2 billion in the third quarter ended Sept. 31, a 11.4% increase from the same time a year earlier.

Alibaba Group Holding Ltd., one of JD.com’s top rivals in China, also announced plans to launch a ChatGPT rival this week. Alibaba is currently testing the chatbot internally. It could reportedly be integrated into several of the company’s applications after launch.

Many of the advanced AI chatbots that have debuted in recent quarters, including OpenAI’s ChatGPT, are based on neural network design known as a Transformer architecture. It was developed by Google LLC researchers in 2017. The primary innovation in the architecture is a feature known as a self-attention mechanism, which enables neural networks to answer more complex questions than was previously possible.

AI models consider a large number of data points when making a decision. To determine whether the word “bank” refers to a financial institution or the terrain along a river, an AI model might examine the sentence in which the word is used. Neural networks also consult contextual information when performing other tasks.

Thanks to their self-attention mechanism, Transformer models such as ChatGPT can understand contextual data better than earlier AI software. The mechanism prioritizes data points based on their relevance to the task at hand. This allows AI models to base decisions on the most important contextual details, which leads to increased accuracy. 

Some earlier neural network designs also included features for prioritizing contextual information. However, they had technical limitations that caused them to lose some of this information during processing. Transformer models can retain all the relevant contextual data about the task they are tackling, which allows them to generate more accurate results.

Google, whose researchers developed the Transformer architecture in 2017, has also announced plans to launch a ChatGPT rival recently. The company’s chatbot is known as Bard and can generate detailed natural language answers to search queries. Google will start making Bard broadly available to users in the coming weeks. 

Image: Flickr