The Pentagon allegedly tested OpenAI models via Microsoft Azure while OpenAI’s military ban was still active, raising significant AI ethics concerns. Sources cited by Wired reveal that this workaround exploited Microsoft’s $13 billion investment in OpenAI, allowing access to GPT models before the company officially lifted its military prohibition in January 2024. This situation poses critical questions about policy enforcement when technology is licensed through third-party partners, potentially leading to unauthorized use. The timing is notable as OpenAI now seeks defense contracts, suggesting that military access to its technology predates the company’s public policy shift. OpenAI has long championed responsible AI development and maintaining a military use ban, but this revelation indicates a complex reality of AI governance. The collaboration with Microsoft provided the Pentagon a loophole to experiment with OpenAI’s models, illustrating the challenges in aligning corporate and ethical standards in the deployment of artificial intelligence.
Source link
