Amidst the escalating US-China trade war, Baidu CEO Robin Li highlighted the release of over 70 large language models (LLMs) by Chinese institutions and technology companies, intensifying the competition for technological supremacy. These models, with over one billion parameters, signify China’s pursuit of excellence in chipmaking, quantum computing, and AI, a trend mirrored by companies like ByteDance, SenseTime, Zhipu AI, MiniMax, and Baichuan Intelligent Technology.
On the other side of the battlefield, US-headquartered institutions and companies have also been actively releasing large language models, including Meta’s Llama 2, OpenAI’s GPT 3.5 and GPT-4, Google’s PaLM 2 and LaMDA, IBM’s watsonx models, Anthorpic’s Claude 2, Databricks’ Dolly, and Amazon’s Titan foundation models. While the exact number is unknown, a Stanford University study revealed that American institutions accounted for the majority (54% in 2022) of the world’s large language and multimodal models.
However, collaboration between Chinese and American organizations in the development of these models has declined significantly since 2020. The increase in US restrictions on chipmaking technology export to China has depleted the once-booming AI research collaborations between the two countries. In fact, the study reported the most negligible year-on-year growth (2.1%) since 2010.
Despite this setback, the study also emphasized the continued US dominance in private investment and the number of newly funded AI companies. In 2022, the US received $47.4 billion in private AI investment, more than three times that of China. Furthermore, the US surpassed both the European Union and the UK in the number of newly funded AI companies by a factor of 1.9 and China by 3.4.
While Chinese firms may gain an advantage in the domestic market due to the availability of large local language datasets, the race for LLM and AI supremacy remains uncertain. Analysts recognize China’s rapid progress, with companies like Baidu and Alibaba driving innovation. China has also made significant strides in AI research papers and citations, particularly in the education sector, though it lags behind in other sectors.
As the US and China continue to clash in the trade war arena, US lawmakers have urged the Biden administration to take further action to impede China’s advancement in areas like artificial intelligence and quantum computing. With strategic moves like trade restrictions on Chinese memory chip maker Changxin Memory Technologies, both nations are vying for dominance in these critical technological fields.
FAQ
Q: How many large language models has China released?
A: Chinese institutions and technology companies have released over 70 large language models with over one billion parameters.
Q: Which companies are competing in the AI space?
A: Companies like Baidu, ByteDance, SenseTime, Zhipu AI, MiniMax, and Baichuan Intelligent Technology are competing in the AI space.
Q: How has collaboration between China and the US changed?
A: Collaboration between Chinese and American organizations for developing large language models has declined since 2020 due to increased US restrictions on chipmaking technology export.
Q: Who leads in private AI investment and the number of newly funded AI companies?
A: The US leads in both private AI investment and the number of newly funded AI companies, surpassing China by a significant margin.
Q: Is China catching up to the US in AI research?
A: China shows progress in AI research, particularly in the education sector, although it still lags behind in non-profit, government, and industry sectors.
Q: How are US lawmakers responding to China’s advancements?
A: US lawmakers urge the Biden administration to take further action, such as trade restrictions, to impede China’s progress in areas like artificial intelligence and quantum computing.