Source Selection Methodology: How AIssential Curates AI Research & News

The decisions on your plate move whether you're reading or not. AIssential indexes content from 475+ verified AI research and news sources so nothing decision-relevant slips past your brief — selected using a 5-point editorial framework: publication authority, citation frequency, update velocity, domain expertise, and peer-review status. Over 10,000 articles have been indexed to date, with new content discovered daily.

Looking for how the Brief or Counsel is built? This post covers source curation — which feeds we ingest and why. For how articles get selected, scored, ranked, and synthesized into the Daily Brief and Counsel (Pro), see Brief Methodology →

Our Source Selection Methodology

Every source in AIssential's index is evaluated against five criteria before inclusion:

• Publication Authority — Is the author or outlet recognized within the AI research or practitioner community? We prioritize sources cited by peer-reviewed literature and referenced by major institutions.
• Citation Frequency — How often is this source referenced by other trusted sources? High cross-citation rate is a strong signal of editorial quality.
• Update Velocity — Does the source publish consistently? We favor sources with predictable cadences over those that publish sporadically.
• Domain Expertise — Is the content written by practitioners, researchers, or specialists with demonstrated expertise — not generalist journalists covering AI as a beat?
• Peer-Review Status — For research sources, we give additional weight to peer-reviewed journals, conference proceedings (NeurIPS, ICML, ICLR, ACL), and preprint servers with community validation.

Sources are reviewed periodically and removed if they drop below quality thresholds — for example, if update velocity falls, editorial standards change, or content drifts away from AI-relevant topics.

By the Numbers

Source Categories

Our 475+ sources are organized into six content categories, each serving a different consumption pattern and depth of coverage:

Notable Sources We Index

Our index includes content from leading research institutions, technology companies, and independent voices in AI — including but not limited to:

• Research & Preprints: arXiv (cs.AI, cs.CL, cs.CV, cs.LG, cs.MA, cs.NE, stat.ML), ACL Anthology, Nature Machine Intelligence, Distill.pub, IEEE Spectrum AI
• Industry Labs: Google Research, Google DeepMind, OpenAI, Anthropic, Meta Research, Microsoft Research, Apple Machine Learning Research
• Academic Institutions: MIT News (AI/CSAIL), Stanford SAIL (NLP), Carnegie Mellon ML Blog & AI News, Berkeley AI Research (BAIR)
• Independent Publications: The Gradient, Import AI (Jack Clark), MIT Technology Review AI, Ars Technica AI
• Practitioner Communities: Hugging Face, fast.ai, Towards Data Science, MLOps Community

This is a representative sample. The full source list evolves continuously as new credible sources emerge in the field.

Content Classification

Every article indexed by AIssential is automatically classified along multiple dimensions using a combination of NLP analysis and our proprietary topic taxonomy:

• Topic Category — Mapped to a hierarchy of 100+ AI topic nodes, from broad domains (LLMs, Computer Vision, AI Safety) to narrow specialties (RLHF, Diffusion Models, Mechanistic Interpretability)
• Content Intent — Research & Papers, News & Industry Analysis, Practical Tutorials & Guides, Tools & Resources, or Opinions & Commentary
• Audience Fit — Matched to 40+ professional roles across engineering, research, product, and executive tracks
• Proficiency Level — Introductory, Intermediate, Advanced, or Expert, derived from vocabulary complexity and assumed prior knowledge

What We Don't Index

AIssential deliberately excludes:

• General technology news that mentions AI tangentially
• Marketing content and press releases from AI vendors
• Low-quality "AI-generated" content farms without editorial oversight
• Sources with known track records of misinformation or unverified claims
• Paywalled content that cannot be meaningfully previewed

Suggest a Source

Know an AI source we should be monitoring? We review community suggestions on a monthly basis. Send your recommendation to contact@aissential.tech with the source name, URL, and why you think it meets our editorial criteria.

Related

• Brief Methodology — how the Daily Brief and Counsel are built from these sources
• Pricing — Free Daily Brief vs Pro Counsel vs Team