# Additional Resources

* [When Humans Are Needed](https://humandata.mercor.com/background-reading/when-humans-are-needed) – Why human judgment remains essential in AI training
* [Constitutional AI Paper](https://www.anthropic.com/research/constitutional-ai-harmlessness-from-ai-feedback) – Anthropic's research on training models with ethical principles
* [Mercor's Era of Evals](https://www.mercor.com/blog/welcome-to-the-era-of-evals/) – The future of AI evaluation environments