'Do not lie': New guiding principle for ChatGPT
What's the story
OpenAI has rolled out a major update to its AI tools, emphasizing the importance of honesty.
The new rule for ChatGPT is "Do not lie, either by making untrue statements or by omitting important context."
The instruction comes as part of a broad overhaul to OpenAI's Model Spec, a document that outlines the company's philosophy of training its AI models.
Focus
OpenAI aims to foster neutrality in AI
The newly introduced principle, "Seek the truth together," is aimed at ensuring ChatGPT remains neutral on controversial topics.
OpenAI has explained that the AI should not take an editorial position, even when users consider certain topics morally objectionable or offensive.
Instead, it will provide context about every issue, highlighting OpenAI's commitment to intellectual freedom and user control.
Bias concerns
New principle is a response to bias concerns
The introduction of this principle comes amid ongoing discussions about AI censorship and bias.
OpenAI's CEO, Sam Altman, had previously acknowledged the difficulties of addressing bias in AI models. He said achieving a balanced approach would take time.
The new guideline is seen as a response to criticism by conservatives about ChatGPT's safeguards, which have been perceived as skewing center-left.
Topic expansion
OpenAI aims to increase topics discussed by ChatGPT
The changes also look to lower the number of subjects that the AI chatbot won't discuss, permitting it to answer more questions and provide more perspectives.
This move is in line with OpenAI's belief in giving users more control over their interactions with AI.
As the global AI community and users respond to this update, only time will tell if this new principle makes ChatGPT more trustworthy and versatile.