OpenAI releases new guidelines to assess 'catastrophic' AI risks

SAN FRANCISCO - OpenAI, the creator of ChatGPT, has unveiled its latest guidelines for assessing "catastrophic risks" posed by artificial intelligence in the models currently in development, reported Xinhua.

In a "Preparedness Framework" released on Monday, the company said, "We believe the scientific study of catastrophic risks from AI has fallen far short of where we need to be."

The framework is designed to "help address this gap", it said.

The monitoring and evaluation team, established in October, will concentrate on evaluating "frontier models" currently in development with capabilities surpassing those of the most advanced AI software.

The team will analyse each new model, assigning a risk level ranging from "low" to "critical" in four primary categories.

Once the risks are identified, they will be forwarded to OpenAI's Safety Advisory Group. The head of OpenAI will then decide if changes should be made to the AI model to reduce the risks.

The board of directors will be kept informed and have the authority to overturn management decisions, the company said. - BERNAMA