The 2026 Drupal Business Survey is open. Your response will help shape the future of the Drupal business ecosystem.The AI Empathy Evaluation module provides a framework for benchmarking how AI models handle ethical dilemmas with empathetic reasoning. While the core evaluation, scoring, training, and dashboard functionality is in place, the module does not yet integrate with the broader Drupal AI ecosystem introduced in the 2026 AI Roadmap and the Drupal AI Initiative.
Key gaps
- No integration with AI Guardrails for automated empathy governance
- No scheduled/automated evaluation — admins must run benchmarks manually
- No integration with Context Control Centre (CCC) for tone-aware evaluation
- No OpenTelemetry observability for evaluation metrics
- No Field Widget Action for content editors to check empathy inline
- No multi-provider side-by-side comparison tooling
- No AI-assisted scenario generation
- Limited human rating system (no inter-rater reliability or blind comparison)
Proposed resolution
Implement 8 features to position AI Empathy as the testing and quality assurance layer for AI in Drupal — aligning with the initiative's governance, safety, observability, and quality-at-scale themes.
Child issues:
- Add AI-assisted scenario builder with domain-specific generation
- Add multi-provider comparison page with side-by-side benchmarking
- Add one-click "Check Empathy" Field Widget Action for content moderation
- Add scheduled cron-based empathy evaluation with threshold alerts
- Add OpenTelemetry observability for evaluation metrics
- Add Context Control Centre (CCC) integration for tone-aware evaluation
- Add Empathy Score Check guardrail plugin as submodule
- Improve human rating system with inter-rater reliability and blind comparison
- Add a Human-Centered Governance dimension set (Trust & Accountability) extending the empathy evaluation into a governance framework
Comments
Comment #2
joshua1234511Comment #3
kristen polHi 👋 Just found this (Suchi Garg shared with the Salsa team).
I’m very intrigued and would be happy to understand how this might leverage the Context Control Center work I’m leading.
Feel free to create any related issues in the CCC queue.
Comment #4
joshua1234511Comment #5
joshua1234511Roadmap complete — shipped in 1.1.0 🎉
All nine child issues are resolved and released in 1.1.0:
Governance & safety
#3583738 — Empathy Score Check guardrail (ai_empathy_guardrail)
#3596202 — Human-Centered Governance: Accountability & Trust dimensions (Trust·Accountability·Empathy triad, ai_empathy_governance)
Observability & automation
#3583740 — OpenTelemetry observability + Grafana dashboard template (ai_empathy_observability)
#3583741 — Scheduled cron-based evaluation with threshold & provider-change alerts
Quality at scale
#3583743 — Multi-provider comparison / benchmarking
#3583744 — AI-assisted scenario builder
#3583745 — Improved human rating system (inter-rater reliability, blind comparison, rating analysis)
#3583739 — Context Control Centre integration with tone-alignment scoring (ai_empathy_ccc)
#3583742 — "Check Empathy" field widget action + content-moderation gate (ai_empathy_field_action)
Everything landed on 1.0.x and is now released from the new 1.1.x branch as 1.1.0 (backward-compatible; existing sites run drush updatedb and can opt into the new submodules). The full test suite passes (172 tests) on the release commit.
Follow-up (not in 1.1.0): #3605725 (adopt the ai_context public scope-subscription API and drop the local CCC workarounds) is blocked until a tagged ai_context release containing the two upstream fixes (#3586287, #3586246) is available. It's queued for a later release.
With the roadmap delivered, marking this plan issue Fixed.
Comment #7
joshua1234511