The Digital Press

Thousands of contract workers, hired by Google through subcontractors like GlobalLogic, perform content rating and moderation to improve AI outputs, but face high pressure, poor pay, and limited transparency about their work.

Why it matters: These human raters are essential in training safe AI responses but often suffer mental health issues and job insecurity.

The big picture: AI models rely heavily on underpaid, overworked humans behind the scenes, revealing a hidden labor layer in AI development.

The stakes: Loosening content moderation guidelines risks harmful AI outputs, while workers struggle with unrealistic deadlines and inadequate support.

Commenters say: Readers debate if conditions are truly exploitative or typical contract work, with some downplaying concerns as exaggerated or insufficiently detailed.