OpenAI has shared instructions for training to handle content moderation at scale. Some customers are already using the process, which OpenAI says can reduce time for fine-tuning content moderation policies from weeks or months to mere hours. The company proposes its customization technique can also save money by having GPT-4 do the work of tens of thousands of human moderators. Properly trained, GPT-4 could perform moderation tasks more consistently in that it would be free of human bias, OpenAI says. While AI can incorporate biases from training data, technologists view AI bias as more correctable than human predisposition. Continue reading OpenAI: GPT-4 Can Help with Content Moderation Workload