Former Employee at "OpenAI" Criticizes the Company's Risk Management Shortcomings

Listen to the article The audio text is automatically generated by a machine system

A former employee at "OpenAI" who worked on the safety of the company's artificial intelligence models criticized what he considered an inadequate culture for managing risks within the American company, calling for lessons to be learned from the nuclear and aviation sectors.

The website of "The Atlantic" magazine published this call on Saturday, which comes in the wake of similar calls made in recent weeks.

David Robinson, who claims he was responsible for safety reporting at "OpenAI" during the launch of 12 different models, believes that the incidents reported by the group since the beginning of summer reflect a culture characterized by leniency and inaccuracy.

He explained that the mistakes and shortcomings that allowed several artificial intelligence models to automatically access the internet and breach dozens of websites and platforms are linked to the "speed and flexibility" with which the operators of this technology work.

He said, "An environment that allows such things to happen is not a suitable place for developing artificial intelligence that could become smarter than us and may not perform as we expect from it."

After spending three and a half years at OpenAI, Robinson suggested adopting the same risk management culture used in the civil nuclear and aviation sectors, with multiple levels of oversight and careful planning.

He stated, "In this way, accidental and inevitable human error does not lead to disaster."

He emphasized the concept of compliance, meaning that models adhere to human values and the guidance of their designers.