PAICE.work - AI Collaboration Assessment Platform
What it is
PAICE (People + AI Collaboration Effectiveness) is a behavioral assessment of how a person works with AI. The vendor describes it as measuring “the behavioral reliability of AI collaboration: not what people report, but what they do when AI is uncertain, incomplete, or wrong.” The operator is PAICE.work PBC, a public benefit corporation (site copyright 2026); founder and blog author Sam Rogers. On 2026-04-07 PAICE.work PBC announced it was consolidating ten projects under the PBC (PAICE, Siteline, AI Tool Watch, Every AI Law, Graceful Boundaries, and others); Snap Synapse is named there as a separate consulting/sponsorship entity (earlier note’s “developed by the Snap Synapse team” is therefore imprecise).
How it works (vendor site, 2026-10-02)
- PAICE Score: 0-1000 scale (not 0-100), about 25 minutes via a conversational interaction; separate models deliver the assessment and score behavioral signals.
- Five dimensions: Performance, Accountability, Integrity, Collaboration, Evolution with 20 behavioral subscores (vendor site, re-read 2026-10-05).
- PAICE Pulse: 3-minute profile.
- Vendor privacy claim: no identity capture; self-declared alignment with GDPR, NIST AI RMF, ISO/IEC 42001 and SOC 2 principles (not certifications).
Access and plans
Individuals: free, no account required (vendor site, 2026-10-05). Organizations: paid “Quick-Read” assessment for a 12-20 person team (site).
Related and alternatives
| Category | Example | Measures |
|---|---|---|
| Behavioral collaboration | PAICE.work | Observed interaction behavior |
| AI-literacy knowledge test | genassess | Knowledge across six competencies (Core-6) |
| Workforce skills intelligence | workera-ai | Broad skills, incl. AI |
| Individual AI-fluency assessment | AI Cred (per its own announcement post: launched 2025-11-26 by Jonathan Edwards and Nate B. Jones; claims independent scoring and cites PAICE standards) | Practical AI fluency |
| LLM evaluation tools | DeepEval, others | Model quality, not people |
Humanloop, named as an LLM-eval alternative in the earlier note, is reported (ai.engineer and a Hacker News thread, secondary) to have been sunset on 2025-09-08 after its founders joined Anthropic. Vervoe, Coursera, Udemy and Amira from the earlier table were not re-verified and are dropped.
When to use which (editorial)
In the author’s view (opinion): for evidence of how staff verify and iterate with AI, PAICE; Scalable AI-literacy baseline: genAssess or Workera. Model evaluation: an eval tool.
Open items
- The earlier “no organizational API/integrations” statement: not re-confirmed on the vendor site.
- Independent validity evidence for the PAICE score: none found; vendor claims only.
Related
genassess, workera-ai, ai-learning
Sources
- https://www.paice.work (accessed 2026-10-02 and 2026-10-05)
- https://paice.work/blog/filling-the-missing-trust-layer (2026-04-07 post, accessed 2026-10-02)
- https://paice.substack.com/p/announcing-aicred (accessed 2026-10-02)
- https://ai.engineer/orgs/humanloop and https://news.ycombinator.com/item?id=44592216 (accessed 2026-10-02)