AI: National Dreams, Global Impact?

“The more AI development becomes a competition between nations, the harder it becomes to ensure the safety of this powerful technology”

China’s “Next Generation Artificial Intelligence Development Plan” sets specific targets for AI development in three stages: by 2020, 2025, and 2030, aiming to achieve breakthroughs in AI technology, become a global AI powerhouse, and ultimately lead the world in AI theory, technology, and applications.  While DeepSeek R1 made headlines, it's probably just one example of a broader trend emerging.

•  “Six Tigers”—Stepfun, Zhipu, Minimax, Moonshot, 01.AI and Baichuan—are considered to be at the forefront of China’s AI sector. Late last year, there were mentions about Chinese firms moving away from building “pre-trained” models that costs billions of dollars to “product” development  (referred as “The Monetization Dilemma”).  Alongside Six Tigers, DeepSeek and ModelBest continue to grow in influence through research. Minimax and Moonshot, are building consumer-facing applications on top of others’ models.

Experts believe, and it appears that the US excels in “short-chain innovation”. i.e. the rapid transformation of technical breakthroughs (Research) into commercial products. Davos had conversations on “Research to Deployment” time. Science-Driven R&D: A strong emphasis on fundamental research, with close ties between academia (leading universities) and industry (companies like Google, Meta, OpenAI).  Cloud Infrastructure: Stronghold: AWS, Azure, and GCP hold approximately 66% of the global cloud market share. While the top three Chinese cloud providers (Alibaba Cloud, Huawei Cloud, and Tencent Cloud) dominate the Chinese market with a combined 70% share, they represent only 8% of the global market. Software Dominance: A mature software ecosystem and expertise in algorithm development, allowing for quick deployment and scaling of AI models. 

•  On the other hand, China's approach is characterized by “long-chain” innovation: the complex, multi-step process of developing and manufacturing phygital products.... Manufacturing Prowess: A world-leading manufacturing base and sophisticated supply chains....Hardware Integration: A growing ability to integrate AI software with hardware, creating intelligent devices and systems.... Industry-Specific Applications: A strong focus on deploying AI in specific industries, leveraging its manufacturing capabilities....Rapid Deployment: A willingness to rapidly deploy and iterate on AI solutions in real-world settings, even if they are not yet perfect....  A Unique, Localized Ecosystem: The restrictions on access to many global AI platforms have fostered a distinct, rapidly evolving, and increasingly self-sufficient AI ecosystem within China. 

AI Supremacy fight is likely to continue in three areas... Hardware: American firms account for 71.5% of the industry’s global market cap. Its leadership stems from its expertise in chip design, intellectual property (IP), and software tools.... Models: The US has OpenAI (GPT-4), Anthropic (Claude 3), and allies like France's Mistral. China counters with Baidu (Ernie Bot), Alibaba (Qwen), DeepSeek, and Kimi.ai..... Consumption (Open Source Vs. Proprietary): The open-source movement (Meta's Llama 3, Mistral, DeepSeek) is challenging the dominance of proprietary models, creating a more diverse and globally accessible AI landscape.

Uniting for AI: A Global Partnership: The Paris AI Action Summit held this week shifted from abstract discussions to tangible steps, highlighting the need for collaboration among government, industry, and academia to create a reliable AI environment. This environment rests on three core elements: scientific research, practical applications, and established norms.  The summit acknowledged the challenges of governing AI globally, promoting open-source collaboration while maintaining security, and bridging the gap between inclusivity and geopolitical differences.  

• Open-source AI emerging as a major theme, with participants stressing its potential to enhance safety, promote democratic values, and challenge dominance. Open-source ecosystems were lauded for their ability to drive collaboration and innovation while ensuring accessibility and transparency. Current AI, a foundation launched by President Macron, aims to steer AI towards social benefit, backed by nine governments and $400 million (aiming for $2.5 billion). However, the summit also acknowledged the importance of securing advanced AI models from cyber threats.
• A New, Collaborative AI Bloc:  The summit saw the emergence of a new AI bloc, largely comprised of European nations and others committed to open-source principles. This bloc, representing roughly half the global population (60+ countries signed a non-binding statement), aims to promote “open, inclusive, transparent, ethical, safe, secure and trustworthy AI”.  The United Kingdom and the United States did not sign the declaration.

AI race for sure is now even more intense..........