OpenAI’s Former Security Manager: Artificial Intelligence was Organized Like Nuclear Plants

OpenAI’s former head of security made a strong call for controlling artificial intelligence technology. In recent descriptions, it argued that more stringent procedures should be followed during the release of advanced AI models.

The recommendations of the specialist include:

• Multilayer control systems installation, design to identify any critical error
• Frontier AI models before marketing long lasting testing and evaluation periods
• Inspired by the Nuclear security protocols, and carefully follow the planning process**
• From the current approach based on the self-definition of the sector,** pass to the off-grim surveillance mechanisms

The necessity of such an arrangement framework is due to the rapid development of artificial intelligence technology in recent years and increased concerns against potential risks. A nuclear power plant-like approach expresses a comprehensive control network from detection of system errors to each stage of product inspection.

The description of the former security director brings a new perspective to the discussions that continue to manage artificial intelligence in the technology industry. During the development of advanced models, safety measures and ethical assessments deepen the conversations on what level should be kept in the forefront.