Artificial intelligence (AI) technology development company OpenAI said the GPT-4 generative model it developed could be used by social media companies to carry out the content moderation process more efficiently.

According to a Tech Crunch report on Wednesday, the capability relies on GPT-4 being guided by policies in conducting moderation assessments as well as creating a series of test examples of content that may or may not violate policies. These assessments will then be given labels by policy experts that can be used to improve their policies.

“By examining the differences between GPT-4 judgments and human judgments, policy experts can ask GPT-4 to discover the reasoning behind the labels, analyze ambiguities in policy definitions, resolve confusion, and provide further clarification in appropriate policies,” said OpenAI in a statement.

OpenAI claims that the use of GPT-4 can reduce the content moderation process time for social media companies by several hours and that the model has a better approach compared to models from other developers.

Content moderation is a complex task for social media companies like Meta's parent Facebook and Instagram, which deploy thousands of moderators around the world to block harmful content such as child pornography and images of extreme violence.

AI-based content moderation platforms are nothing new. Previously, Google's Counter Abuse and Jigsaw Technology Team launched Perspective to the public several years ago.

Apart from that, several other technology companies are also developing similar platforms such as Spectrum Labs, Cinder, Hive and Oterlu.