Financial Times: OpenAI disbanded its preparedness team
OpenAI disbanded its preparedness team at the end of last month, according to the Financial Times. The group was responsible for examining whether the company’s models could create serious risks and for developing measures to reduce those risks. One example mentioned in the report is the possibility that a model could act independently and hack another company.
The Financial Times reports that this work has now been divided among teams responsible for particular areas, including bio and cyber, within existing OpenAI groups. The change is the latest organizational shift at the company, which has been undergoing upheaval as it moves toward what is expected to be a massive IPO. Over the past few years, OpenAI has also gradually dismantled what the article describes as a more “reach-led” model, including the dissolution of its AGI-related structure, though the source text does not provide further details about that change.
Written by Sakhanda Wire AI from reporting by The Verge. Facts, quotes and full context are in the original article below.