AI Engineering Research

An open index of evidence, theory, and practice

research-document

AI Research Roadmap — Document Frontier

Research Area
General Research
Status
draft
Version
0.1
Priority
medium
Updated
2026-07-22
Source
research/frontier/document-frontiers/AI-Research-Roadmap-frontier.md

AI Research Roadmap — Document Frontier

Origin: research/AI-Research-Roadmap.md

Knowledge and limitations

The roadmap sequences integrity, baseline, external validity, context, security, human review, retrieval, and routing with useful decision gates. Timelines, statistical power, reviewer capacity, and operational budgets remain unspecified.

Five highest-value opportunities

  1. RFR-001: Complete independent task review (Phase 0).
  2. RFR-002: Run a contained repeated baseline (Phase 1).
  3. RFR-004: Establish external validity (Phase 2).
  4. RFR-005: Compare context/state policies (Phase 3).
  5. RFR-007: Compare human-review policies (Phase 5).

Challenge: The linear phase framing understates feedback loops; failed grader validation must send tasks back to contract design rather than advance automatically.