AI Training Costs and Labor Risks Rise for Businesses
Specialized human trainers are increasingly critical for model accuracy, yet firms face rising costs and privacy lawsuits.
Updated on Oct. 20, 2026 in Employment
Live Poll
Do you trust artificial intelligence companies to provide stable, secure employment for workers?
Artificial intelligence labs are increasingly hiring subject matter experts to train large-scale models, with pay rates reaching $400 an hour for specialized roles. This shift reflects a growing demand for high-level professional input to verify engineering and legal outputs, even as contractors face instability and privacy litigation.
Why it matters
Developers require high-quality, expert-validated training data to improve model accuracy and test outputs. This reliance on human expertise is transforming the labor market, though it introduces significant legal and operational risks related to data security for firms hiring these contractors.
AI labs pay an average of $110 per hour for training tasks, with top rates for civil rights attorneys reaching $400 per hour. Meanwhile, firms face legal scrutiny, including seven class action lawsuits filed in April 2026 alleging data breaches.
The players
Mercor
An AI platform facilitating specialized labor and model training, currently valued at a potential $20 billion.
Nvidia
A leading semiconductor and AI infrastructure firm exploring a funding round investment in Mercor.
Deeptune
An AI-focused technology firm acquired by Mercor in July 2026.
Jack Weis
A contractor whose experience with contract termination highlights the instability of the current AI labor market.
The details
Subject matter experts create training material by writing complex problems and grading model-generated responses to ensure accuracy. While this human-in-the-loop strategy aims to reduce errors in AI outputs, some platforms face allegations of excessive monitoring and data privacy failures. These operational risks are surfacing as companies like Mercor scale rapidly toward a potential $20 billion valuation.
Timeline
January 2026: Jack Weis began Mercor contract work.
March 2026: Mercor terminated the contract of Jack Weis as part of an 810-person group.
April 2026: Seven class action lawsuits were filed against Mercor.
July 9, 2026: Mercor announced the acquisition of Deeptune.
Market Landscape
The aggressive hiring of expert trainers tracks with the rise of the data annotator role, which LinkedIn identified as the fourth-fastest-growing U.S. job for 2026. This trend marks a shift toward highly technical human-in-the-loop validation as companies move beyond general data labeling.
Businesses integrating AI should audit their vendors' data-handling practices to mitigate legal risks stemming from potential training data breaches. Operators hiring specialized contractors must balance the need for high-level expertise against the risk of rapid contract turnover and employment litigation.
The takeaway
The demand for niche professional expertise to refine AI outputs is creating a premium labor tier while simultaneously exposing contractors to job volatility. Managers should track contract termination clauses and data privacy liabilities when engaging AI-focused service providers.
Further reading
For more on the current labor market shifts, see our Employment section.
Source note: This article includes information reported by PYMNTS.
Live Poll
Do you trust artificial intelligence companies to provide stable, secure employment for workers?
