The recent launch of the free, open source artificial intelligence model DeepSeek brings benefits and risks to federal agencies seeking to adopt the technology to achieve their missions, according to Dave Vennergrund, vice president of AI and data insights at General Dynamics Information Technology.
Cost-Effective AI
On the one hand, DeepSeek, whose performance matches if not exceeds that of its Western counterparts despite costing less to build and operate, shows that agencies can gain access to more cost-effective and mission-specific AI tools, Vennergrund said in an opinion piece published Friday.
AI Bias & Cybersecurity Issues
On the other hand, because DeepSeek’s developer, Chinese startup DeepThink, trained the AI on data that was curated in line with censorship policies, the AI’s output may reflect certain biases. Also, because the developer was forced to use older chips for the AI due to export controls against China, additional code had to be introduced. The code was meant to optimize the performance of the hardware but it may also pose a cybersecurity threat.
Precautionary Measures
Consequently, government users of DeepSeek and other similar innovations must be alert to possible output inaccuracies. Agencies should also perform cybersecurity assessments in line with prevailing federal supply chain and cybersecurity guidelines. Also, in light of national security and data privacy concerns, government users should refrain from sharing sensitive information with DeepSeek.
Changing the GenAI Landscape
“DeepSeek has changed the generative AI landscape for the moment, but it has not eliminated the need for secure, foundational AI models,” Vennergrund said, adding, “As the demand and use cases for AI continue to expand rapidly, this latest development underscores the opportunity for more cost-effective, mission-specific AI approaches. It also reinforces the necessity for agencies to develop their AI roadmaps and evaluate their AI investments with security, compliance, and mission impact in mind.”