Alibaba's Tongyi Qianwen Language Model Goes Public in China
Summary:
Alibaba's Tongyi Qianwen, a large language model, will be accessible to the public and enterprises in China from September 13. The model, similar to OpenAI's ChatGPT, has been trained on a mix of English and Chinese text. Alibaba Cloud has also launched open-source language models that can comprehend images, texts, and facilitate multi-round question answering in English and Chinese. This expansion comes after the Chinese government's relaxation of restrictions on AI technologies and the introduction of new guidelines for public release. Other Chinese companies have also received approval to launch their AI models.
Alibaba has announced that Tongyi Qianwen, its artificial intelligence system, will be available for public and enterprise access in China starting from September 13. The large language model, similar to ChatGPT, has been trained on a combination of English and Chinese text. Although the exact specifications of the model are unknown, there were rumors that it could have up to 10 trillion parameters, surpassing OpenAI's GPT4. However, these rumors have not been verified. In the past, Alibaba has released two 7-billion parameter models based on the Tongyi Qianwen architecture as open source. In addition, Alibaba Cloud has launched two open-source large vision language models, Qwen-VL and Qwen-VL-Chat, which have the capability to comprehend images, texts, and bounding boxes, and facilitate multi-round question answering in both English and Chinese.
Initially, Tongyi Qianwen was only accessible to a limited group of users during its beta testing phase. However, with the recent relaxation of restrictions on artificial intelligence technologies in China, the public rollout of Tongyi Qianwen has become possible. In June, the Chinese government published guidelines mandating that all AI technologies made available to the public must go through a special vetting and certification process. These regulations came into effect on August 15. Notably, several Chinese companies such as Baidu, Tencent, TikTok, and ByteDance have already received approval to launch their models.
The updated restrictions include provisions that prohibit the generation of images resembling the Chinese president, X Jinping. Additionally, organizations must address objectionable content within a three-month period. Previous versions of the legislation had proposed monetary fines, but this plan has been discarded. While China is exploring the relaxation of its AI regulations, the United States has taken only preliminary measures in regulating these technologies. Recently, Senate Minority Leader Chuck Schumer hosted a forum with various top U.S. tech CEOs and founders to discuss potential policy ideas, marking the first of nine scheduled forums.
Published At
9/13/2023 8:15:00 PM
Disclaimer: Algoine does not endorse any content or product on this page. Readers should conduct their own research before taking any actions related to the asset, company, or any information in this article and assume full responsibility for their decisions. This article should not be considered as investment advice. Our news is prepared with AI support.
Do you suspect this content may be misleading, incomplete, or inappropriate in any way, requiring modification or removal?
We appreciate your report.