OpenAI is reportedly working on a new project, named Strawberry, that will bring advanced thinking capabilities to its AI models.
Reuters reported that the AI will be able to “navigate the Internet autonomously” to conduct “deep research.” Details are said to be kept strictly confidential, even within OpenAI, although Strawberry is believed to be a rebranded version of OpenAI's Q* project, which Reuters reported last year.
In a statement to Reuters, OpenAI says: “We want our AI models to see and understand the world as we do. “Continuous research into new AI capabilities is common practice in this industry, with the common belief that these systems will get better at thinking over time.”
Earlier this week, Bloomberg reported that OpenAI had created a model with “human-like” thinking skills. A demo of this model was reportedly shared with OpenAI employees during an internal all-employee meeting on Tuesday; However, it is unknown if this model is also Strawberry.
All of this comes as OpenAI is under fire for reportedly preventing its employees from reporting risks the company's technology may pose to humanity. A group of employees recently sent a seven-page letter detailing concerns to the Securities and Exchange Commission (SEC) earlier this month.