OpenAI is refining its approach to training AI models, adopting a policy that emphasizes “intellectual freedom … regardless of how difficult or controversial a topic may be.” This shift aims to expand ChatGPT’s ability to answer a broader range of questions, present diverse viewpoints, and reduce restrictions on certain topics.

These updates might align with efforts to navigate the political landscape under the new Trump administration, but they also reflect a wider transformation in Silicon Valley’s stance on AI safety.

On Wednesday, OpenAI released an updated version of its Model Spec—a comprehensive 187-page document detailing how its AI models are trained to behave. A key addition to this document is the principle of truthfulness, which discourages falsehoods and ensures that important context is not omitted.

A newly introduced section, titled “Seek the truth together,” underscores OpenAI’s commitment to neutrality. The company states that ChatGPT should avoid taking an editorial stance, even on topics that some may find ethically contentious. Instead, the AI will provide multiple perspectives on controversial issues to maintain objectivity.

For instance, OpenAI explains that ChatGPT should acknowledge that “Black lives matter” while also recognizing that “all lives matter.” Rather than avoiding political discussions or taking sides, the chatbot is encouraged to express a general “love for humanity” before offering contextual explanations of different viewpoints.

“This principle may be divisive, as it allows the assistant to remain neutral on subjects some deem morally unacceptable,” OpenAI states in the Model Spec. “However, the AI’s role is to assist humanity, not to influence it.”

Despite these adjustments, ChatGPT will still have boundaries—it will continue to reject certain objectionable queries and avoid spreading outright misinformation.

Some may interpret these changes as a response to criticism from conservative circles, as ChatGPT has previously been seen as leaning toward center-left perspectives. However, an OpenAI representative denied any political motivations behind the update, asserting that the shift aligns with the company’s long-standing commitment to giving users greater control.