OpenAI has significantly tightened its internal security following reports that DeepSeek may have replicated its models. DeepSeek allegedly used distillation techniques to launch a competing product earlier this year, prompting a swift response.
OpenAI has introduced strict access protocols to prevent information leaks, including fingerprint scans, offline servers, and a policy restricting internet use without approval. Sensitive projects such as its AI o1 model are now discussed only by approved staff within designated areas.
The company has also boosted cybersecurity staffing and reinforced its data centre defences. Confidential development information is now shielded through 'information tenting'.
These actions coincide with OpenAI's $30 billion deal with Oracle to lease 4.5 gigawatts of data centre capacity across the United States. The partnership plays a central role in OpenAI's growing Stargate infrastructure strategy.
