Alibaba shows off ChatGPT alternative
Companies will be able to build their AI apps using Alibaba's LLM
Chinese technology giant Alibaba has unveiled Tongyi Qianwen, a ChatGPT-style AI chatbot, which the company intends to incorporate into all of its business applications in the near future.
Tongyi Qianwen's official website describes the application as "an efficiency assistant and idea-generator."
Alibaba says Tongyi Qianwen can operate in both English and Chinese, and plans to first integrate it into workspace messaging application DingTalk. That will empower users to accomplish tasks like summarising meeting notes, composing emails and drafting business proposals.
It will also build Tongyi Qianwen into Tmall Genie, Alibaba's smart voice assistant, similar to Amazon's Alexa and Apple's Siri.
"We are at a technological watershed moment driven by generative AI and cloud computing, and businesses across all sectors have started to embrace intelligence transformation to stay ahead of the game," said CEO Daniel Zhang.
Alibaba's research institute, DAMO Academy - now part of its independent cloud computing division - was behind the development of Tongyi Qianwen.
The chatbot is currently only available to corporate clients and a select few media outlets, and then by invitation only.
Alibaba has so far provided little information about either the requirements to employ Tongyi Qianwen, or the associated costs. The company has also not shared any information about the infrastructure behind the model, or its related cloud services.
Tongyi Qianwen's official website currently prompts visitors to provide their email address and phone number to request an invitation code.
Alibaba has, however, said that it intends to use Tongyi Qianwen has a foundation for custom large language models (LLMs). The company plans for customers to use the chatbot to create their own AI applications, without needing to build and train their own LLM.
A custom version of the service is already in beta for some Chinese organisations, and developers can access a beta version of the Tongyi Qianwen API.
The year AI goes mainstream
Increase in the global interest in generative AI, which uses past data to generate new content, has spiked since the launch of ChatGPT in 2022. In recent months, various Chinese companies have announced or hinted at their own AI models and chatbots.
Last month search giant Baidu unveiled Ernie Bot, based off its deep learning model, Ernie - short for "Enhanced Representation through Knowledge Integration".
Access to the Ernie bot was originally limited to a select group of users with invitation codes, but companies can now incorporate the bot into their products by submitting an application.
Chinese AI firm SenseTime launched several new AI products just this week, including a chatbot named SenseChat.
According to Zhang, Alibaba and other companies developing AI models are at the "starting line" of a new phenomenon.
"Seizing this opportunity is a common wish for all of us," he said.
China's cyberspace regulator the CAC (Cyberspace Administration of China) today unveiled draft measures for managing generative AI.
The proposed rules say companies will be accountable for the legitimacy of data they use to train LLMs.
The public can provide feedback on the proposals until 10th May.
Earlier this month, Italy's data protection authority temporarily banned the use of ChatGPT over the collection of users' personal data.
The ban was prompted by a data breach that exposed the personal information of ChatGPT users, including their names, addresses, credit card types and expiration dates.
The regulator also criticised OpenAI's failure to implement a filter to prevent children under 13 from using the service, exposing them to inappropriate content.