<ul data-eligibleForWebStory="false">OpenAI has 'overhauled' its security operations to protect its AI algorithms from corporate espionage.The company is limiting access to new algorithms, implementing tighter restrictions on sensitive data and strengthening staff vetting processes.OpenAI safeguards its model weights with strict policies, including limiting internet connections and increasing physical security at data centers.Microsoft security researchers suggested that individuals linked to DeepSeek may have been exfiltrating data through OpenAI's API.