The Pentagon is discussing plans to create secure environments for generative AI companies to train models on classified data. This will enable AI models like Anthropic’s Claude to become more accurate in military tasks.
Generative AI Companies and Classified Data
The Pentagon is considering allowing generative AI companies to train military-specific models on classified information. This could increase accuracy in tasks such as target analysis but also involves security risks as sensitive information could be integrated into the models.
Training AI models on classified data could result in sensitive information becoming accessible to unauthorized users. Therefore, the Pentagon plans to evaluate the effectiveness of models trained on unclassified data before implementing this new approach.
Source: technologyreview.com
Read also: Nemotron 3 Nano 4B: Compact Hybrid Model for Local AI
