In February, the US President signed an order to develop an AI Action Plan to maintain and strengthen America's dominance in AI. That same month, the Office of Technology and Science issued a call for public feedback. Among others, Anthropic and OpenAI shared their views. Below is a brief summary of their proposals.
A federal program should be established to build power plants with a total capacity of 50 Gigawatts of energy by 2027 specifically for the AI industry. For reference, total average consumption in the USA is around 440 GW (although there is a reserve to produce significantly more), and New York City consumes an average of 6 GW.
Currently, the sale of the most powerful H100/B100 GPUs to China is prohibited, but truncated versions, such as the H20, can be sold and then used to train AI. Anthropic suggests banning even these and making agreements with other countries to monitor smuggling.
Identify areas where AI can already be integrated and start doing so (improving taxation processes, healthcare, etc. – mainly for document processing). Additionally, AI assistants need to be systematically integrated into workflows.
Increase cybersecurity in advanced laboratories by sharing experience with relevant agencies and collaboration with the Five Eyes, an anglosphere intelligence alliance. There should also be a procedure for accessing certain information regarding model development.
Create federal testing infrastructure to enhance the US government's capabilities to quickly assess AI models in terms of national security threats.
Monitor the economic impact of AI (where it is used and what results it yields – this will help estimate tax implications and create employment models for the future).
Link the conscientious use (copyright) of data to national security and state that if China continues to have free access to data while American companies do not due to lawsuits and delays, the AI race is effectively over.
Suggest banning the use of Chinese LLMs, such as DeepSeek, if there are suspicions they violate privacy and pose security threats.
Launch and encourage initiatives with significant investments in AI infrastructure, for example, inviting Sovereign Funds or creating economic zones with accelerated approval processes.
Modernize accreditation processes for federal procurements so AI tools can pass through faster and more easily (especially new iterations/models when the old ones are already in use – not starting from scratch each time).
Export restrictions for China, expansion of the energy network, introducing a National Transmission Highway Act to accelerate laying of communication/power lines, integrating AI into government structures, and closer collaboration with various agencies to further train models on their data (for better performance upon implementation).
Boost Personal Productivity With AI Automation on Latenode
Use each model mentioned above – ChatGPT, Claude, and DeepSeek – to boost your workflows and productivity. Do it right now with AI automation on Latenode without restrictions! Simplify time management, data processing, database management, email summarization, and lots of other tasks! Check out a few simple, fully customizable solutions on Latenode:
Integrate voice commands with AI models to generate content for social media posts, emails, or any other text-based tasks. Inside the template, you’ll get a free tutorial on how to work with Telegram Bots.
Stay updated on the weather effortlessly, without having to check multiple sources.
Imagine what you could achieve with the right tools at your fingertips. With Latenode, you're getting this productivity boost at a fair, credit-based pricing model. Are you ready to maximize your potential? The future of work is here, and it's more accessible than ever.
Don't Miss Out! Sign Up Now and start your 14-day free trial to experience the future of AI-driven productivity.
.png)
.svg){kind=icon}
