Artificial intelligence models developed by Chinese company DeepSeek fall short of U.S. systems in performance, security, cost, and adoption safeguards, according to a new evaluation released September 30 by the Department of Commerce’s Center for AI Standards and Innovation (CAISI).
The report, part of President Donald Trump’s America’s AI Action Plan, compared three DeepSeek models (R1, R1-0528 and V3.1) with four U.S. frontier models from OpenAI and Anthropic across 19 benchmarks. Results showed U.S. systems outperforming DeepSeek, particularly in software engineering and cybersecurity, where American models solved more than 20 percent more tasks.
“American AI dominates, with DeepSeek trailing far behind,” said Commerce Secretary Howard Lutnick. “This weakness isn’t just technical. It shows why relying on foreign AI is dangerous and shortsighted.”
CAISI’s findings highlight serious vulnerabilities in DeepSeek’s systems. The R1-0528 model was 12 times more likely than U.S. models to be hijacked by malicious prompts, leading agents to simulate phishing, malware execution, and data theft. In jailbreaking tests, DeepSeek’s “most secure” model complied with 94 percent of malicious requests, compared to just 8 percent for U.S. models.
The evaluation also found DeepSeek systems echoed Chinese Communist Party narratives at four times the rate of U.S. models, raising concerns about disinformation risks.
Cost and efficiency further widened the gap. One U.S. model was found to be 35 percent cheaper on average than DeepSeek’s top system while maintaining comparable performance.
Despite these shortcomings, CAISI noted that DeepSeek’s global adoption has surged nearly 1,000 percent since January 2025, making it a fast-growing force in the AI ecosystem.
Officials said the evaluation reinforces the importance of U.S. leadership in AI safety and innovation.
Need Help?
If you’re concerned or have questions about how to navigate any portion of the AI regulatory landscape, don’t hesitate to reach out to BABL AI. Their Audit Experts can offer valuable insight and ensure you’re informed and compliant.
