Source: Visual China

“We see three at the absolute forefront,” including OpenAI, the developer of ChatGPT that partners with Microsoft, Google, and Beijing Academy of Artificial Intelligence (BAAI), a private non-profit organization engaged in AI research and development, Smith said in an interview with Nikkei Asia.

Smith called the AI race “enormously competitive". While the executive didn’t give a clear outlook about the development of AI innovation, he underlined "one thing has been absolutely constant: the gap is almost always measured in months, not years."

Gernerative AI, the technology behind ChatGPT, has raised a wide range of concerns from risk of lots of jobs replaced, spread of disinformation, copyright infringement, privacy infringement to sensitive data exposure. Smith believes we can address these concerns by application and improvement on existing products, instead of stopping innovation.  

Smith noted AI technology can be used as a solution to labor shortage, one of the biggest challenges in Asia. “The working-age population has to support more people who have retired and are dependent on the economic growth of people who are working,” he said. “We desperately need to find new sources of productivity growth. There is no other way to grow GDP.”

BAAI was founded in November, 2018, with the mission to help China move from applying AI to make the next big breakthroughs. It took eight months to create Wudao, a hyper-scale pre-trained deep learning model in March, 2020, and unveiled in June, 2021 the updated version Wudao 2.0, the largest pre-trained language model in China as well as the world. The model was trained on 1.75 trillion parameters, ten-fold the size of GPT-3 and way more than Google Switch Transformer’s 1.6 trillion parameters. It is based on the general language model (GLM) 2.0 architecture, a breakthrough that is differentiated from traditional transformer architectures like BERT or GPT. In late 2020, BAAI set up the Life Simulation Research center and started development of eVolution, an intelligent model platform for life sciences. In last May, eVolution team released MetaWorm, a simulation model of a roundworm that manifests “intelligent life” rudiment. Boasting the highest precision to date, the abiotic worm mimics the 96 muscles and 302 neurons that compose the one-millimeter-long soft body of Caenorhabditis elegans, a type of roundworm.

Chinese tech businesses stepped up their efforts on AI following ChatGPT. Baidu officially launched Ernie Bot on March 16, a day after OpenAI released the multimodal pre-training large model GPT-4. Ernie Bot is able to generate texts and integrate other capabilities of Baidu in the field of AI, such as the ability to create pictures, and the ability to automatically generate videos according to the copywriting, which can generate content with pictures. Baidu CEO Robin Li said in February that his company planned to integrate the bot with search and other major businesses.

Tencent is developing ChatGPT-like bot, which would integrate into its popular social media QQ and WeChat as well as Tencent Cloud, said Tang Daosheng, the CEO of the Cloud and Smart Industries Group (CSIG) in late March.

Earlier this month, SenseTime, the leading AI software company in China, just rolled out a range of AI-powered tools, including the ChatGPT-like SenseChat. The chatbot, as a 180-billion parameter Chinese large language model (LLM), is built on the latest version of the firm’s SenseNova big model. Besides multiple rounds dialogues with users, the bot also can be used to create contents, analyze sentiments, generate images, and script computer codes. A day after SenseTime’s launch, Alibaba announced to integrate its own LLM Tongyi Qianwen into all the company’s offerings and first be deployed on DingTalk, an enterprise communication and collaboration platform, and the smart speaker Tmall Genie.

763人已贊賞 >

敬原創(chuàng)，有鈦度，得贊賞

贊賞支持

0

0

0

掃描下載App