Halfway through 2026, the remote AI training market looks different than it did even six months ago. The commodity end (basic annotation, simple preference labeling) has softened as models improve, while the expert end — domain specialists, coders, scientists, credentialed professionals — has grown sharply. Here's a grounded snapshot of where things stand this summer: what's hiring, where pay is going, and the trends worth acting on.
The headline: expertise is the moat
The single biggest shift is the flight to expertise. As base models got good at the easy stuff, the human work that still moves the needle is increasingly specialized: evaluating graduate-level reasoning, catching subtle domain errors, and producing gold-standard answers a model can't yet generate. That's why the best-paying roles this summer are for people with real, verifiable credentials — and why generic "no-skill" gigs pay less than they did a year ago.
What's hiring most right now
- Specialized ML / research — the fastest-growing category. Frontier labs are sourcing narrow expertise (diffusion, world models, interpretability, computational science) at $150–$200/hr. See our data scientists & ML guide.
- Credentialed professionals — law, medicine, finance, accounting. Deep demand and stable rates ($75–$200/hr).
- Software engineering — evergreen; code evaluation and secure-code review keep growing.
- Multilingual & audio — steady demand as models expand language coverage; strong for non-English speakers.
- Generalist evaluation — still hiring, but the largest supply and the softest rates.
Where pay is heading
Bifurcating. The top tier (frontier-lab specialists) is holding or rising — $150–$200+/hr is now common for narrow ML and senior professional work. The bottom tier (generic microtasks) is flat to down as automation eats the simplest tasks. The practical takeaway: the return on positioning yourself as a specialist has never been higher. Our full pay breakdown has the tier-by-tier numbers.
Platform landscape this summer
- Mercor and AfterQuery — the expert-tier leaders; AfterQuery's public catalog has expanded past 100 specialized roles.
- micro1 — still the broadest catalog and best first stop for most people.
- Mindrift — the largest global freelance pool by volume.
- Turing and Handshake AI — engineering-heavy and academic-fellowship niches respectively.
The full ranked view is in our best-platforms guide and featured-platforms roundup.
Three trends to act on
- Specialize your profile. A single, precise domain out-earns a broad "generalist" profile by a wide margin now. The getting-accepted guide covers positioning.
- Run multiple platforms. Demand is bursty; the top earners keep 2–3 platforms active and take the best-paying work at any moment.
- Move up the value chain. If you're doing commodity annotation, use it to build a track record, then pivot to your professional specialty where the rates are climbing.
The honest caveat
This is a snapshot, not a forecast. The market moves fast — task categories that are hot this summer can cool as models improve, and new specializations open constantly. The durable strategy isn't chasing any single trend; it's leading with verifiable expertise and staying on multiple platforms. And the fundamentals still hold: it's legitimate remote contract work (watch for scams), paid like 1099 income (taxes guide).
Get started
Browse what's live right now on the home page, filter by your specialty, and read the complete AI training jobs guide for the full playbook.