Some former TV writers and showrunners are now working paid gigs training and testing AI, with one account reporting rates as high as $150 per hour. A Wired first-person report describes writers doing timestamp annotation, face spotting, furniture labeling and red-team testing for companies the writer names as Mercor, Outlier, Task-ify, Turing, Handshake and Micro1. An AOL feature documented the industry squeeze that pushed many toward nontraditional work, citing a Writers Guild report that counted 1,819 TV writing jobs in 2023-24 versus 3,011 in 2018-19. Hiring cycles are ongoing and piecemeal, so the flow of entertainment professionals into AI operations appears likely to continue for now.

The people who used to write and run TV shows are now doing the back-end labor that helps build and police generative AI. That shift comes through in a first-person Wired account by a writer who says she began contracting to train AI systems in September 2025, and who names a half-dozen companies she worked for.

What the work looks like

The Wired writer lays out a set of operational tasks that read more like data-processing than creative writing. Contractors are paid to annotate video timestamps down to the millisecond, judge chatbot tone, identify faces in group photos and flag strangers for removal, label patterns in furniture, and run so-called red-team scenarios that intentionally try to elicit unsafe or illicit model outputs so safety systems can be tested.

Those assignments require attention to timing, detail and narrative sense, the writer says, but employers treat the labor as data work rather than an extension of television storytelling. The account describes a mix of short tasks and longer labeling jobs, filled on a contract-by-contract basis rather than through steady staff positions.

Reported compensation varies by contract. The Wired writer describes being paid roughly $150 per hour on some assignments. She also says landing those gigs often meant passing multiple unpaid tests, completing lengthy application processes with as many as ten applications or long trial tasks, and sometimes undergoing interviews conducted by automated recruiting agents.

Why writers are taking the jobs

The shift is tied to shrinking opportunity in TV. An AOL feature on Hollywood labor and career pressure details how staffing in television has contracted since the prior peak, and it cites a Writers Guild report that counted 1,819 TV writing jobs in the 2023-24 season compared with 3,011 in 2018-19.

Mastheads

That kind of scarcity, the AOL piece notes, pushes experienced writers to seek any paid work they can find.

The Wired writer describes joining online groups where peers posted urgent financial needs and tips for side work. The AOL reporting adds that competition for open staff positions can be extreme, with one hiring process drawing 400 submissions for two slots. So writers and other entertainment professionals are treating AI-labeling gigs as a stopgap or income source while staff jobs remain scarce.

The labels and judgments these contractors provide are adjacent to storytelling skills, but many in the industry say the roles are operational. The Wired account emphasizes that seasoned narrative professionals supply judgment about timing, tone and plausibility, while companies treat the tasks as fungible data work that can be parceled out to dispersed contractors.

That combination creates what both pieces frame as a new labor pipeline from entertainment into AI operations. Experienced creatives supply a form of narrative expertise. Startups and vendors supply demand for labeled data and safety testing. The result is a dispersed, sometimes opaque workforce doing high-volume training and safety labor for AI systems.

Not every detail in the first-person account is independently corroborated. The Wired piece provides the specific company names, the $150-per-hour pay figure, and granular descriptions of onboarding like ten applications and 20 unpaid hours of testing. Those particulars appear only in the Wired account and aren't verified in the AOL feature. Three other sources reviewed by the research editor were archival items unrelated to Hollywood labor or AI training and offered no independent confirmation for the Wired claims.

Still, both Wired and AOL converge on the central point that career pressure in television is pushing talent toward alternate forms of paid work. The Wired writer describes piecemeal hiring patterns rather than any single coordinated rollout, and the AOL feature documents the broader industry context that makes such transitions more likely.

For the contractors themselves, the work comes with trade-offs. Some of the tasks reproduce audition-like submission cycles and demand that applicants present what employers want to hear, according to AOL. The intermittent nature of contract work can also complicate access to guild-covered benefits and the stability that staff positions once provided.

Startups and vendors hiring for labeling and safety work have not, in the reporting, announced a public initiative to hire entertainment professionals en masse. Instead, the Wired account describes ongoing, piecemeal recruitment through online groups and platform listings. That pattern suggests the flow of writers into AI operations will continue as long as demand for annotated data and safety testing remains high and television staff opportunities remain limited.

Related Articles

The writer says her first AI-training contract began recently, and she describes continued, piecemeal hiring cycles rather than a single industry program.

This article was created with AI assistance.