Hollywood's Hidden AI Workforce: Screenwriters Describe Surge in Secretive Training Contracts as Traditional TV Work Dries Up

Insiders in the entertainment industry are describing a quiet but accelerating shift: workers who once built television content are now being recruited to train the AI systems reshaping their own field. According to one screenwriter who spoke to Wired, the work has become so prevalent that it now functions as a fallback income source for creatives struggling to find traditional employment. In eight months, this individual completed 20 contracts across five separate platforms, describing the experience as fundamentally different from creative labor. The pattern suggests that AI training work has become a structural feature of Hollywood's labor market, filling gaps left by contraction in scripted television production.

The platforms involved remain largely undisclosed by workers, who report operating under non-disclosure agreements that limit what they can reveal about the contracts. This secrecy itself has become a defining characteristic of the arrangement. Workers describe the labor as repetitive, low-prestige, and disconnected from the creative work they originally pursued. The comparison to waiting tables—a job associated with impermanence and transition—underscores how workers themselves perceive this shift: not as a new career path, but as survival mode in an industry whose traditional structures are under pressure. The five platforms referenced represent a cross-section of companies building AI capabilities, suggesting the practice is distributed across the sector rather than concentrated with a single buyer.

The implications extend beyond individual experience. Each contract represents a decision by entertainment-industry platforms to source training data from workers with firsthand knowledge of creative content, potentially embedding industry-specific knowledge into systems that could eventually perform that creative work at scale. Labor advocates have warned that this dynamic creates a self-reinforcing cycle: displaced workers train the systems that may further displace them. The question of whether these platforms are investing in transition support, reskilling pathways, or long-term creative workforces remains largely unaddressed in public disclosures. For now, the contracts continue, and workers with deep expertise in scripted content are earning income by teaching machines to mimic it.