Scale AI cuts more contractors as it shifts toward more specialized AI training (www.businessinsider.com)

Scale AI this week cut an entire Dallas contractor team — the New Projects Organization (NPO) of a dozen-plus workers focused on “generalist” tasks like rewriting chatbot responses — and offered four weeks’ severance plus healthcare through October. Laid-off staff were invited to join Scale’s gig platform, Outlier, which connects thousands of freelancers to model-training tasks. Scale confirmed the move as the wind-down of a small experimental onsite program and said it affects only a fraction of its workforce and not customer delivery. The move underscores a broader technical shift in AI data work: as base chatbots improve, demand is moving away from generalist annotation and rewriting toward expert, domain-specific labeling (medicine, robotics, finance) and higher-skill adversarial testing. The change follows turbulence since Meta’s $14.3B stake in Scale, which prompted some clients to pause work and preceded earlier cuts (14% of staff and hundreds of contractors) and the recent Red Team reductions. Other AI players, like xAI, have made similar pivots away from generalist tutor roles. For practitioners and vendors, the signal is clear: model-training labor markets are trending toward specialized, higher-skill contributions and flexible gig channels rather than broad, onsite generalist teams.