Alibaba announced that its proprietary large language model, an artificial intelligence system called “Unified Qianwen”, will be available to the public and businesses nationwide starting on September 13.
Uni-President Qianwen is a large language model similar to ChatGPT, trained on English and Chinese text corpora. While its exact specifications are unclear—early rumors suggested it would be trained on up to 10 trillion parameters, 10 times more than OpenAI’s GPT4, but these remain unconfirmed—Alibaba has previously released two An open-source model with 100 million parameters unifies Qian Wen’s architecture.
Alibaba Cloud has launched two open source large visual language models, Qwen-VL and Qwen-VL-Chat, which can understand images, text and bounding boxes in prompts, and support multiple rounds of question and answer in Chinese and English.
Learn more: https://t.co/LrVNzNIRVy… pic.twitter.com/0t0Et6j1BU
— Alibaba Cloud Intelligent Business Group (@cloudintelgrp) August 30, 2023
Previously, Uni-President Qianwen was only open to a limited user group during the testing phase. The public rollout coincides with the People’s Republic of China’s recent easing of restrictions related to the use of artificial intelligence (AI) technology.
A set of guidelines issued by the Chinese government in June stipulated that all artificial intelligence technology released to the public going forward would require special review and certification processes.
The regulation takes effect on August 15. According to Cointelegraph, several Chinese companies have been approved to launch models, including Baidu, Tencent, TikTok and ByteDance.
related: Tencent launches ChatGPT rival in China as US ban on AI chips continues
The updated restrictions include a ban on generating images of Chinese President Xi Jinping and a requirement that organizations address objectionable content within three months. Previous versions of the legislation called for fines, but as Cointelegraph reported, those plans were scrapped.
As China explores deregulation, the United States has taken only preliminary steps to regulate artificial intelligence technology. Most recently on September 13, Senate Minority Leader Chuck Schumer hosted a number of the nation’s top tech CEOs and founders in the first of nine scheduled forums to discuss potential policy ideas.