The company is revamping how ChatGPT responds to controversial topics, explicitly embracing “intellectual freedom” and neutrality in a move that could reshape AI content moderation.
The company’s updated Model Spec, a 187-page guide on how its AI should behave, introduces a key principle: ChatGPT should not lie, either by stating falsehoods or omitting important context.
The policy shift means ChatGPT will provide more perspectives on complex issues rather than declining to answer or favoring one viewpoint.
OpenAI is aming for its AI to remain neutral, even on topics that some consider morally or politically charged.
For example, ChatGPT should acknowledge both “Black lives matter” and “all lives matter,” offering historical and social context rather than choosing a stance. The company frames this as a commitment to assisting humanity rather than shaping public opinion.
This change follows conservative criticism that ChatGPT’s safeguards skew left. A viral incident where ChatGPT refused to write a poem praising Donald Trump—while doing so for Joe Biden—sparked accusations of bias. OpenAI CEO Sam Altman acknowledged the issue at the time, calling it a “shortcoming” the company was working to fix. However, OpenAI denies that its recent updates are an effort to appease the Trump administration, instead describing them as part of a long-held belief in user control.
The policy shift also aligns with a wider trend in Silicon Valley. Over the past year, major tech firms have retreated from progressive initiatives, scaling back diversity programs and content moderation efforts. Elon Musk’s xAI positioned its Grok chatbot as a free speech alternative to OpenAI, and Meta’s Mark Zuckerberg recently refocused the company around First Amendment principles.
OpenAI’s latest update removes ChatGPT’s warning messages when users violate its policies, though the company says this is purely cosmetic. While the AI will still reject certain harmful requests, the overall goal is to make interactions feel less restricted.
This shift comes at a pivotal moment for OpenAI. As it pursues its ambitious Stargate project—a US$500 billion AI data center—the company’s relationship with Washington will be critical. At the same time, ChatGPT is vying to challenge Google as the go-to source for information. OpenAI’s new stance suggests it sees open discourse as the best way to navigate both political and technological challenges.
