Skip to content

Further Reading

Background sources for journalists covering FCC or the broader algorithmic-accountability beat.

FCC primary sources

  • The README and CLAUDE.md files in the main repository are the fastest way to get a summary of the current state of the project. Read them first.
  • The 5 Architecture Decision Records (ADRs) under docs/decisions/ explain the rationale behind the plugin system, persona schema, workflow graphs, compliance framework, and evidence graph. Each is short (2-4 pages) and readable.
  • The 24-chapter FCC Guidebook under docs/guidebook/ is a systematic treatment. Chapters 1, 14, 20, 22 are the most relevant for journalism.
  • The three-book series under docs/books/ (Understanding FCC, Building with FCC, Advanced FCC) provides a deeper and narrative treatment.
  • The release notes (RELEASE_NOTES.md) record what changed in each version.

Sister-project repositories

  • POLARIS — policy-ontology reasoning. Search for the POLARIS repo on GitHub under the same owner.
  • LYRA — graph-backed UX. Same.
  • Ten constellation-verticals — each independent:
  • Ophiuchus (finance)
  • Serpens (health)
  • Libra (justice)
  • Crater (legal records)
  • Scutum (insurance)
  • Norma (metrology)
  • Pyxis (logistics)
  • Vela (energy)
  • Columba (communications)
  • Caelum (environmental)

Regulatory primary sources

  • EU AI Act (Regulation 2024/1689). Official text on EUR-Lex. Articles 5-7 define risk; Annex III lists high-risk use cases. About 140 pages; skimmable.
  • NIST AI RMF 1.0 (January 2023). Publication available from NIST. Organized into four functions (Govern, Map, Measure, Manage) with subcategories.
  • ISO/IEC 42001:2023. AI management system standard. Behind an ISO paywall; summaries are widely available.
  • OECD AI Principles (2019, updated 2024). OECD website. High-level but influential.

Academic literature

Classics:

  • Mitchell, M. et al. (2019). Model Cards for Model Reporting. FAT '19.
  • Gebru, T. et al. (2018). Datasheets for Datasets.
  • Raji, I. D. et al. (2020). Closing the AI Accountability Gap.

Multi-agent systems:

  • Park, J. et al. (2023). Generative Agents. Useful background on agent simulation, though differently motivated than FCC.
  • Wooldridge, M. (2009). An Introduction to MultiAgent Systems. A classic textbook.

Auditability and governance:

  • Raji, I. D. et al. (2022). Outsider Oversight: Designing a Third-party Audit Ecosystem for AI Governance.
  • Costanza-Chock, S., Raji, I. D., Buolamwini, J. (2022). Who Audits the Auditors?

Policy and civil-society reports

  • Brookings AI Governance series.
  • Stanford HAI Policy Briefs.
  • AlgorithmWatch (EU-focused, civil-society watchdog).
  • Ada Lovelace Institute (UK-focused).
  • Center for Democracy & Technology (US-focused).

These are often underused by working journalists. Their reports are accessible, quotable, and generally fair.

Journalistic precedent

Useful pieces to read for style and structure (search by title):

  • Coverage of the COMPAS recidivism tool (ProPublica, 2016) — template for algorithmic-accountability reporting.
  • Coverage of the Dutch childcare-benefits scandal (various Dutch outlets, 2018-2021) — template for government-AI malfunction reporting.
  • Coverage of the Apple Card credit-limit allegations (Bloomberg et al., 2019) — template for consumer-AI fairness reporting.

Not all are FCC-relevant in content, but their methods — archive-and-quote, request-and-file, reproduce-when-possible — map cleanly onto FCC-based stories.

Broadcast and podcast resources

  • The Gradient Podcast. Deep technical but widely listened.
  • Your Undivided Attention. Accessible ethics/tech.
  • The TWIML AI Podcast. Longer-form technical.

Useful for pre-interview warming when you are new to the beat. Not a substitute for primary sources.

Code and data repositories beyond FCC

  • Papers With Code. Track performance claims in published AI papers.
  • Hugging Face. Model cards for thousands of models, useful for comparison.
  • OpenReview. Peer review transcripts for many AI papers, invaluable for understanding how a result was received.

Glossaries

  • FCC's own Ecosystem Glossary (yes, the policy-makers track's glossary is equally useful for journalism).
  • Stanford HAI AI Index (annual). Has a good glossary and trendlines.

First story (you have never covered FCC):

  1. This track's Getting Started (20 min).
  2. The FCC README (10 min).
  3. One ADR — start with ADR-0004 Compliance Framework (30 min).
  4. This track's Explaining FCC to the Public (30 min).
  5. File.

Second story (you have covered once):

  1. FCC Guidebook chapters 14 and 20 (1-2 hours).
  2. Mitchell et al. 2019 (45 min).
  3. One ADR you have not read.
  4. One constellation-vertical README relevant to the beat.

Ongoing coverage:

  1. Subscribe to FCC release notifications.
  2. Maintain a source list across builders, operators, regulators, affected users.
  3. Re-read one ADR per quarter.
  4. Keep a running list of promised-but-unshared artifacts (it becomes a story).

Where next

Return to Citation Guide to tighten attribution before filing, or to Interview Personas to prep sources. Back to the track index.