After FBI arrested a Saudi National for hiding a spy device at OpenAI, employees said ‘not surprised’: Controls are so lenient that …

Former OpenAI employees say they were not surprised after the company discovered a hidden audio recording device in an executive conference room in October, as they had raised concerns about its internal security practices for years. According to a Reuters report, some employees warned OpenAI's leadership and board that its security controls were too weak and could allow outsiders, including foreign governments, to access sensitive information about its artificial intelligence (AI) technology. The concerns came into focus after FBI announced on October 13 that a Saudi national, Abdulrahman Alfozan, 35, had been arrested and charged with allegedly placing a hidden recording device in an OpenAI executive conference room. Alfozan was working as an engineer at the company at the time.OpenAI employees had raised security concerns for yearsAccording to five people familiar with the matter cited by Reuters, OpenAI employees raised concerns about the company's security practices even before ChatGPT launched in late 2022. Some employees believed that the company's internal controls were so weak that foreign governments or other outside actors could easily obtain sensitive details about its technology. They also said that employees with access to OpenAI's code could view details of projects that were not directly related to their work.Some staff members reportedly felt that OpenAI's leadership was more focused on keeping up with rivals such as Google and Anthropic than on strengthening security.What did the Saudi national allegedly doAccording to a court filing cited by Reuters, Abdulrahman Alfozan allegedly used an AI tool designed for OpenAI employees to search for internal documents about executives travelling to Saudi Arabia. He also reportedly looked for information about which OpenAI officials were likely to attend meetings with Saudi ministers.The filing also stated that Alfozan was in regular contact with a Saudi royal who had tried to recruit him for a new AI venture in Saudi Arabia.OpenAI has started strengthening its security systemsOpenAI has reportedly begun taking steps to address its internal security concerns. Thibault Sottiaux, a senior executive overseeing product development, has recently taken charge of an internal effort to secure the company's systems. In August, Sottiaux reassigned around a quarter of OpenAI's production engineers to work on improving its security infrastructure, according to one person familiar with the matter.The company has also restricted access to sensitive work and said that its most sensitive projects are stored in separate code repositories with strict access controls. OpenAI said its main code repository does not contain all of its code or critical details of its AI models, including model weights, which determine how a model behaves. The company also said it now conducts regular security reviews of its internal tools.You use AI every day. Now get your AI Quotient. Take the AIQ test.