"OpenAI stated that its AI systems 'will not be intentionally deployed for domestic surveillance of U.S. citizens and subjects.' The Pentagon confirmed that AI services will also not be used by intelligence agencies such as the NSA. OpenAI drew three explicit red lines: no mass domestic surveillance, no fully autonomous weapons systems, and no automated decisions with high consequences."
"Negotiations between Anthropic and the Pentagon broke down over the use of AI for mass surveillance and fully autonomous weapons systems. The US government subsequently labeled the company a 'supply chain risk.' OpenAI soon emerged as an alternative after Anthropic decided that the AI requirements from the US Department of Defense did not provide enough 'guardrails' for the deployment of the technology."
OpenAI is negotiating to provide AI models to NATO's unclassified networks and has secured a deal with the Pentagon after Anthropic declined due to insufficient safety guardrails. The arrangement follows failed negotiations between Anthropic and the Department of Defense over concerns about mass surveillance and fully autonomous weapons systems. OpenAI established three explicit restrictions: no mass domestic surveillance of U.S. citizens, no fully autonomous weapons deployment, and no automated decisions with significant consequences. CEO Sam Altman acknowledged the deal appeared opportunistic and sloppy initially. AI systems are deployed exclusively through cloud infrastructure with security-cleared personnel involved, and safety mechanisms remain intact for military applications.
#ai-military-deployment #nato-and-pentagon #ai-safety-restrictions #autonomous-weapons #government-ai-policy
Read at Techzine Global
Unable to calculate read time
Collection
[
|
...
]