Is the US losing its edge in the field of AI?

In reports submitted to the U.S. government, the companies warned that America's advantage in the AI ​​field is gradually narrowing as advanced models like China's DeepSeek R1 become increasingly sophisticated.

These reports were submitted in March 2025 in response to a call for input on the government's AI Action Plan.

increasing presenceThe rise of AI in China

DeepSeek R1, an AI model from China, has attracted the attention of American developers. OpenAI sees DeepSeek as proof that the technological gap between the US and China is gradually narrowing.

The company described DeepSeek as "state-subsidized, controlled, and provided free of charge," and expressed concerns about China's potential to shape the development of global AI.

OpenAI even compared DeepSeek to the Chinese telecommunications giant Huawei, warning that Beijing's regulations could allow the government to force DeepSeek to interfere with sensitive systems or critical infrastructure.

Additionally, OpenAI also expressed concerns about data privacy, noting that requests to share information with the Chinese government could enhance the state's surveillance capabilities.

Anthropic, another American AI company focused on biosecurity, emphasized that DeepSeek R1 "can answer most questions about biological weapons, even though it was designed to avoid dangerous content." This contrasts with the stringent safety protocols applied in US-developed AI models.

However, the competition isn't limited to DeepSeek. Baidu, the Chinese search engine giant, recently launched two powerful AI models, the Ernie X1 and Ernie 4.5, to challenge even the most advanced Western systems.

According to Baidu, Ernie X1, a reasoning model, has performance comparable to DeepSeek R1 but at half the cost. Meanwhile, Ernie 4.5 costs only 1% of OpenAI's GPT-4.5 but outperforms it in several evaluation criteria.

Both OpenAI and Anthropic view this race not just as technological but also ideological, a competition between "democratic AI" based on Western values ​​and "authoritarian AI" controlled by the state.

However, the recent successes of Baidu and DeepSeek suggest that cost and accessibility can play a decisive role in the global adoption of AI, transcending ideological differences.

Concerns about AI security and infrastructureAmerica

American technology companies have also expressed concerns about the security and infrastructure challenges associated with AI development.

In its report to the government, OpenAI focused on the risks posed by the Chinese government's influence on models like DeepSeek, while Anthropic highlighted the biosecurity threat posed by AI's increasingly advanced capabilities.

Anthropic revealed that their Claude 3.7 Sonnet model has shown significant improvements in supporting biological weapons development, highlighting the dual nature of advanced AI systems.

At the same time, the company also pointed out loopholes in the US export control policy. Although Nvidia's H20 chips still comply with export regulations, they still have strong performance in text generation, a crucial element in reinforcement learning.

Therefore, Anthropic calls on the government to further tighten controls to prevent China from gaining a technological advantage.

Meanwhile, Google takes a more balanced approach. The company acknowledges security risks but warns against the potential for overregulation.

According to Google, overly strict export controls could weaken the competitiveness of the U.S., especially for domestic cloud service providers and AI development companies.

Instead of imposing broad restrictions, Google proposes targeted controls aimed at protecting national security without hindering the industry's growth.

OpenAI, Anthropic, and Google all emphasized the need for increased government oversight of AI security. Anthropic called for expanding the AI ​​Security Institute and strengthening the role of the National Institute of Standards and Technology (NIST) in assessing and mitigating AI-related threats.

Economic competitiveness and energy demand

Reports submitted to the US government also highlight the role of economic factors in the development of AI. Anthropic warns of infrastructure challenges, predicting that by 2027, training just one advanced AI model could consume up to 5 gigawatts of electricity.

To address this issue, the company proposed building 50 gigawatts of dedicated power capacity for AI and expediting the transmission line approval process.

Meanwhile, recent statements from Baidu highlight the importance of cost-effective AI. Models like Ernie 4.5 and X1 are said to cost only a fraction of what their Western counterparts do, with significantly lower token processing costs compared to OpenAI's current models.

This pricing strategy could put pressure on American AI companies to lower costs to maintain a competitive edge. OpenAI describes this as an ideological competition between Western and Chinese AI, arguing that a free-market approach would foster innovation and deliver greater benefits to consumers.

In its reports, Google focuses on more practical policy recommendations. The company urges the federal government to increase investment in AI research, expand access to government contracts for technology companies, and simplify export controls to support the industry's growth.

Management strategiesWHO

A unified strategy for AI governance has become a common thread in all three submissions. OpenAI proposes a regulatory framework overseen by the Department of Commerce, warning that fragmented regulations across states could accelerate the trend of AI development overseas.

This company also advocates for a tiered export control model, allowing democratic nations broader access to American AI technology while restricting access for other regimes.

Meanwhile, Anthropic called for tighter export controls on AI hardware and training data, emphasizing that even small improvements in model performance could give China a significant strategic advantage.

Google takes a different approach, focusing on copyright and intellectual property issues. The company argues that current policies, based on the principle of "fair use," are fundamental to AI development and warns that overly strict copyright regulations could undermine the competitiveness of American companies against their Chinese counterparts.

All three companies emphasized the need for faster AI deployment in the public sector. OpenAI proposed removing administrative barriers to AI testing and procurement, and supported Anthropic's campaign to simplify technology procurement processes from federal agencies.

Google also agrees with these reforms, emphasizing the importance of improving interoperability within government cloud infrastructure to more effectively drive AI development.

Maintaining a competitive advantage

The submissions from OpenAI, Anthropic, and Google all reflect a shared concern about protecting America's leadership in AI, especially as China increases competitive pressure.

The rise of DeepSeek R1, along with advanced models from Baidu, poses not only technological challenges but also puts pressure on costs and accessibility.

As AI continues to develop at a breakneck pace, the U.S. government will face the challenge of balancing national security, economic growth, and maintaining its leading position in the global technology race. This will likely remain one of the most important policy challenges in the coming years.