The unequal treatment of demographic groups by ChatGPT/OpenAI content moderation

Re:The unequal treatment of demographic groups by ChatGPT/Op... by ineptech on Thursday February 02, 2023 @12:41PM (#34630217)

There's something richly ironic about bashing your political outgroup in a thread discussing why an LLM trained on internet posts would exhibit political bias.

[ Parent | on HN ]

Re:The unequal treatment of demographic groups by ChatGPT/Op... by flutas on Thursday February 02, 2023 @02:17PM (#34631680)

> why an LLM trained

The article isn't even about ChatGPT, it's about the moderation endpoint that OpenAI provides.

https://platform.openai.com/docs/guides/moderation

[ Parent | on HN ]

Re:The unequal treatment of demographic groups by ChatGPT/Op... by ineptech on Thursday February 02, 2023 @02:27PM (#34631816)

And that endpoint is powered by what, if not an LLM?

[ Parent | on HN ]

Re:The unequal treatment of demographic groups by ChatGPT/Op... by erichocean on Thursday February 02, 2023 @06:14PM (#34634663)

> why an LLM trained on internet posts would exhibit political bias

You're moving the goalposts. The LLM doing moderation is trained on human responses to Internet posts—literally human censorship training an LLM to introduce bias at scale.

[ Parent | on HN ]