Artificial intelligence and machine learning research

Field at a glance

Frontier AI Safety & Governance

• “Concrete Problems in AI Safety”: The 2016 Paper Explained

  What the 2016 paper “Concrete Problems in AI Safety” actually says: its five named problems — avoiding side effects, avoiding reward hacking, scalable…

• AI Accountability: Who's Responsible When an AI System Causes Harm?

  What AI accountability means as a general concept, the mechanisms (roles, sign-offs, audit trails, incident-reporting obligations) that implement it,…

• AI Control Evaluations: The Evidentiary Method Redwood Research Built

  Most frontier AI evaluations ask what a model can do, or what it tends to do. A control evaluation asks a third question: if the model were actively t…

• AI Red Teaming: Definition and How It Works in Frontier AI Safety

  AI red teaming inside a frontier safety framework is a safeguard-testing commitment, not the AppSec/jailbreak service most search results describe --…

• AI Safety vs. AI Security: What the Distinction Actually Means

  A short definitional guide distinguishing AI safety (preventing an AI system from causing unintended harm through its own behavior or capabilities) fr…

• Bletchley Declaration vs Seoul Declaration: The International AI Safety Summit ProcessComparison

  Bletchley (Nov 2023) and Seoul (May 2024) are the first two outputs of the international AI safety summit process. They aren't symmetrical: Seoul spli…

• Building an AI Safety Framework with NIST's Govern, Map, Measure, Manage Functions

  A walkthrough of NIST AI RMF's Govern, Map, Measure, and Manage functions as a methodology for building an AI safety framework from scratch, with a pr…

• Building an Internal AI Safety Incident Response Program

  What an internal AI safety incident response program needs structurally: detection channels, triage against a severity taxonomy, escalation to a named…

• Building an Internal Audit Function for Frontier AI Safety

  An internal AI-safety audit function checks, on a schedule, whether safety commitments are actually being met — distinct from incident response, which…

• C2PA Content Credentials: How Provenance Actually Works

  Content Credentials are often described as a nutrition label for digital media. The metaphor oversells them. C2PA is a signed-assertion format with a…

• CAISI's Evaluation Cadence for Open-Weight and PRC-Origin AI Models

  CAISI has published cyber-capability assessments of three PRC-lab open-weight models in four months -- GLM-5.2, Kimi K3 (with UK AISI), and GLM-5.3 --…

• California SB 243 (Companion Chatbots): The State’s Second AI Reporting Channel

  California already has one AI reporting regime — SB 53, which routes critical safety incidents to the Office of Emergency Services. SB 243, signed a f…

• California SB 53 (Transparency in Frontier Artificial Intelligence Act): The Foundational Explainer

  A plain-language breakdown of what California SB 53 actually requires: who counts as a frontier developer, the frontier AI framework and transparency…

• Capability Thresholds: 14 Labs and Regulators, One Undefined Term

  Anthropic, OpenAI, and Google DeepMind all set “capability thresholds.” So do California, the EU, and the Frontier Model Forum. Almost none of them sa…

• Comparing AI Safety Terms Across Frameworks: A NIKOLAI Crosswalk Guide

  How CASRAI's NIKOLAI crosswalk tables show where threshold, evaluation, safeguard, and accountable-decision-maker language actually differs across RSP…

• EFF vs. AI Now Institute: Where They Disagree on AI Safety Framework DesignComparison

  EFF and the AI Now Institute both argue today's AI-lab safety commitments fall short — for opposite reasons. A side-by-side of their diagnoses, eviden…

• EU AI Act vs California SB 53: Frontier Model Transparency Requirements ComparedComparison

  How the EU AI Act's general-purpose AI provisions and California's SB 53 (Transparency in Frontier Artificial Intelligence Act) differ on who is cover…

• EU AI Act vs NIST AI RMF vs ISO/IEC 42001Comparison

  The EU AI Act, the NIST AI RMF, and ISO/IEC 42001 are not three versions of the same thing: one is binding law, one is voluntary US guidance, and one…

• Executive Order 14409: Frontier Models as Cyber Defense Infrastructure

  Executive Order 14409, signed 2 June 2026, is the third Trump AI order and the first to treat frontier models as defensive cyber infrastructure rather…

• Explainable AI (XAI) Explained: Methods, Regulation, and Why It Matters
• Foundation Models vs. FDA's SaMD FrameworkComparison

  FDA's own AI-enabled-medical-devices page admits the agency will need new methods to identify devices built on foundation models and LLMs -- its exist…

• Frontier AI Labs: Who They Are and What Safety Frameworks They Publish

  A directory of the developers generally treated as frontier AI labs under SB 53 and the EU AI Act, and which safety framework each one publishes, with…

• Frontier AI Safety Timeline: 2023-2026

  Frontier AI safety moved from a voluntary international declaration to five overlapping binding statutes in under three years. This page lists every m…

• Gray Swan Arena: Crowdsourced Red-Teaming as Pre-Deployment Evidence

  Gray Swan's Arena crowdsources adversarial testing to 15,000+ red-teamers, and two competitions with UK AISI and US CAISI produced public papers. What…

• How to Grade a Frontier AI Safety Framework: The SaferAI Rubric

  SaferAI, an independent nonprofit, scores frontier AI labs' published safety frameworks against a four-dimension rubric. Here's how the rubric works a…

• HTI-1's 31 Source Attributes: What a Certified EHR Must Disclose About a Predictive Model

  The most itemised model-disclosure rule in force in the US is not an AI law but a health-IT certification criterion: 45 CFR 170.315(b)(11) requires a…

• Incident Type: The Taxonomy No One Has Actually Published

  SB 53, the EU GPAI Code, OpenAI, Anthropic, and the Frontier Model Forum each classify AI incidents differently — and only one of them names anything…

• Inspect: The Open-Source Evaluation Harness Under Third-Party AI Evals

  Almost every argument about third-party AI evaluation is an argument about access, independence and publication rights. Underneath those arguments sit…

• International AI Safety Report 2026: What It Found

  What the Bengio-chaired International AI Safety Report 2026 found on frontier AI capabilities, malicious-use and control risks, and industry safety fr…

• Is AI Red-Teaming Security Theater? What the Research Says

  Carnegie Mellon researchers surveyed the AI red-teaming literature and analyzed six real exercises -- Bing Chat, GPT-4, Gopher, two Claude releases, a…

• Is AI Safety a Property of the Model? The Narayanan-Kapoor Critique

  Arvind Narayanan and Sayash Kapoor's March 2024 essay 'AI Safety Is Not a Model Property' argues that capability-threshold frameworks -- Anthropic's R…

• Italy's Law 132/2025: A National AI Statute Inside the EU AI Act

  Regulation (EU) 2024/1689 is directly applicable in every member state, so what is left for a national AI law to do? Italy's Law 23 September 2025 n.…

• Machine Unlearning and Algorithmic Disgorgement: Ordering a Model Deleted Is Not the Same as Deleting It

  Since 2019 the US Federal Trade Commission has, at least six times, ordered an organisation not merely to delete data but to destroy the models built…

• Mapping Declarations: How Organizations Verify and Confirm Their Own AI Safety Terminology in NIKOLAI

  Every NIKOLAI crosswalk row starts as an unconfirmed shadow mapping. Here's how an organization verifies its identity and gets a reviewed declaration…

• Mapping Your AI Safety Program to the EU AI Act's GPAI Code of Practice

  Where an existing risk register, governance framework, and incident response program already satisfy the GPAI Code of Practice, and where they need ex…

• Marginal Risk vs. Absolute Risk: The Contested Framing Behind Frontier AI Safety Cases

  Anthropic's RSP rewrote itself around a distinction between a model's absolute risk and its marginal risk relative to what other developers already fi…

• Model Cards, System Cards, and the EU Model Documentation Form

  Model cards, system cards and the EU Model Documentation Form are three different instruments with three different levels of legal force. What each ac…

• NIKOLAI's Track System: A Map of the Frontier AI Safety Landscape (N1-N10)

  Not sure which frontier-AI-safety guide covers your question? NIKOLAI, CASRAI's own dictionary of frontier-AI-safety elements, organizes the whole top…

• NIST AI RMF Generative AI Profile: What AI 600-1 Adds

  NIST AI 600-1, the Generative AI Profile, is a separate July 2024 publication that extends the base AI RMF (NIST AI 100-1) for generative AI specifica…

• NIST AI RMF vs ISO/IEC 42001Comparison

  NIST AI RMF is a free, voluntary framework with no certification; ISO/IEC 42001 is a certifiable management-system standard with real audit costs and…

• NIST's AI Agent Standards Initiative: What It Covers and Its Current Status

  NIST's Center for AI Standards and Innovation launched the AI Agent Standards Initiative in February 2026 to cover AI agent security, identity, and au…

• Open-Weight AI Models: The Safety Case For vs. AgainstComparison

  Once a lab releases a model's weights, the decision cannot be undone. This comparison sets out the actual case for open-weight release (independent sa…

• Responsible Scaling Policy (RSP): What It Is and How the Major Labs Compare

  What a Responsible Scaling Policy is, how capability thresholds and safeguard tiers work, and how Anthropic's RSP compares to OpenAI's Preparedness Fr…

• Safety Cases: The Missing Methodology in Frontier AI Governance

  What a safety case is as a methodology, GovAI's rubric and safety-case papers, how they differ from SaferAI's rubric, and the 2023 paper behind SB 53…

• Seoul Frontier AI Safety Commitments: The Signatory List and What They Pledged

  The 20 companies that have signed the Seoul Frontier AI Safety Commitments, the full list including the February 2025 additions, and how the pledge re…

• Singapore's Model AI Governance Framework for Agentic AI

  Singapore's IMDA and AI Verify Foundation published a voluntary Model AI Governance Framework for Agentic AI (v1.5, May–June 2026) built around four p…

• Singapore's Model AI Governance Framework for Generative AI

  IMDA and the AI Verify Foundation published the Model AI Governance Framework for Generative AI on 19 June 2024 — a voluntary, nine-dimension framewor…

• The AI Safety Index: What It Rates and Who Publishes It

  The Future of Life Institute publishes the AI Safety Index, grading frontier AI companies across six domains. Here's what it rates and how it differs…

• The Algorithmic Accountability Act: What the Federal Bill Would Require

  The Algorithmic Accountability Act is a federal bill, not a law — repeatedly introduced since 2019 and still pending. What it would require, its curre…

• The Center for AI Safety: What It Is and What It Does

  An institutional profile of the Center for AI Safety (CAIS): its mission, the exact text of its 2023 Statement on AI Risk, its safety research and fie…

• The Council of Europe HUDERIA Methodology and Model: What It Asks For, and What It Explicitly Is Not

  HUDERIA's four elements, COBRA's four steps and three contexts, the avoid-mitigate-restore-compensate hierarchy, and the Council of Europe's own state…

• The Frontier Model Forum: What It Is and What It Does

  An institutional profile of the Frontier Model Forum (FMF): its founding members, mission, funding, and how it differs from independent evaluators lik…

• The IDAIS-Beijing Statement, Explained: AI Safety’s Red Lines From China

  The IDAIS-Beijing Consensus Statement on Red Lines, co-signed by Andrew Yao, Zeng Yi and Xue Lan alongside Bengio, Hinton and Russell, is CASRAI’s fir…

• The International AI Safety Institute Network: What It Is, Now NAAIMES

  There's a coordination body connecting the individual national AI safety institutes — CAISI, UK AISI and others — to each other. It was founded in Nov…

• The One Redaction Reason No Lab Will Admit To

  Anthropic, SB 53, Meta, OpenAI, the EU's GPAI Code, and METR each name reasons a frontier AI developer can redact for. Together they cover six categor…

• The Texas Responsible AI Governance Act (TRAIGA): What It Requires

  Texas's TRAIGA (HB 149) takes effect January 1, 2026. Unlike Colorado's law, it bans specific AI uses by intent rather than regulating high-stakes dec…

• The UN Scientific Panel vs. the International AI Safety Report: Which One Is 'the IPCC for AI'?Comparison

  Two international bodies produce scientific assessments of AI risk, both are routinely described as 'the IPCC for AI', and Yoshua Bengio chairs both o…

• UK AI Safety Institute vs AI Security Institute: The 2025 Rename Explained

  The UK government renamed its AI Safety Institute to the AI Security Institute on 14 February 2025. It is the same organisation under a new name, not…

• Untangling the US AI Safety Bills: Stop Rogue AI, the Emergency Button, and Thune-Klobuchar ComparedComparison

  Three US federal AI-safety bills are moving at once in September 2026, and it is easy to conflate them. The House's Stop Rogue AI Act (H.R. 10362) wou…

• Vietnam's Law on Artificial Intelligence No. 134/2025/QH15

  On 10 December 2025 Vietnam's National Assembly passed Law No. 134/2025/QH15 on Artificial Intelligence, 429 votes to 5. It took effect on 1 March 202…

• What Actually Gets Graded: AILuminate vs. the AI Safety Index vs. SaferAIComparison

  Three well-known AI safety scorecards are routinely quoted side by side as though they rank the same things. They do not. MLCommons AILuminate grades…

• What Anthropic Pays for a Jailbreak: Inside Its Model Safety Bug Bounty

  Anthropic will pay up to $15,000 through HackerOne for a novel, universal jailbreak against its CBRN and cybersecurity safeguards. Here's exactly what…

• What Gets Redacted From AI Safety Reports, and Who Has to Say So

  Anthropic, Meta, California SB 53, and METR all require frontier AI developers to disclose when they redact a safety report, and why. The EU's GPAI Co…

• What Is a Frontier AI Model?

  "Frontier AI model" is a capability class, not a specific product or lab. This guide walks through the three definitional approaches in active use — c…

• What Is AI Alignment? Definition and Why It Matters for Frontier AI Safety

  AI alignment is the technical problem of making a model behavior match its developers intended goals and values. This guide defines alignment, disting…

• What Is AI Safety? A Plain-Language Guide

  AI safety is the field working to prevent AI systems from causing harm, whether through misuse, accidents, or loss of control. Here's the working voca…

• What Is METR? How Its AI Safety Evaluations Work

  An institutional profile of METR (Model Evaluation and Threat Research): what it is, what it evaluates, how its evaluations work, and its relationship…

• What Is NIKOLAI? CASRAI's Frontier-AI-Safety Dictionary Explained

  NIKOLAI is CASRAI's own frontier-AI-safety dictionary (nikolai-v0.2): 64 elements across 10 tracks covering thresholds, evaluations, safeguards, incid…

• What Is Responsible AI? Principles, Frameworks, and How to Operationalize Them

  Responsible AI principles — fairness, transparency, accountability, human oversight, security — explained, with how to turn them into a working framew…

• Who Checks the Checkers: Evaluator Independence in AI Safety

  Anthropic's RSP names a financial-interest and personal-relationship test for its external reviewers; the EU's GPAI Code asks only for “adequate quali…

• Who Counts as 'Frontier AI'? Eleven Scope Tests Compared

  The EU AI Act never says ‘frontier model.’ SB 53's ‘catastrophic risk’ isn't the EU's ‘systemic risk.’ An 11-organization crosswalk of the if-then tes…

Research Tools & Software

• ALCC (ASCR Leadership Computing Challenge): Supercomputing Access Guide

  A researcher's guide to the ASCR Leadership Computing Challenge (ALCC), detailing supercomputing allocation caps, PeerNet submission, and pacing rules…

• Choosing and Governing Large Language Models for Research: A Framework

  General AI leaderboards rank intelligence, speed, cost, and context window — not the things that decide whether a model is safe for research work: cit…

• Contextual Bandits: Research Software Guide

  Implement contextual bandits for research using Vowpal Wabbit. Learn multiline ADF formatting, exploration policies, and offline evaluation techniques…

• Embodied AI: Definition, Components and Research Areas

  Embodied AI explained: how two arXiv papers define it, its four components, four research targets, and how it differs from LLM-only systems.…

• GraphCast: Running AI Global Weather Forecasts

  A research guide on deploying GraphCast, Google DeepMind's graph neural network model for medium-range global numerical weather forecasting.…

• How to Request an NSF ACCESS Compute Allocation

  A step-by-step walkthrough of requesting an NSF ACCESS supercomputing allocation: choosing Explore, Discover, Accelerate, or Maximize, what the reques…

• Hugging Face Datasets: Install, Load, Stream, Create

  Install the datasets library, load or stream a Hugging Face dataset, pin its revision for reproducibility, and publish your own with a dataset card.…

• Image Annotation: Types, Free Tools and Export Formats

  Image annotation guide for researchers: annotation types, a Label Studio walkthrough, CVAT as a free alternative, export formats and coordinate pitfal…

• Image Recognition: How It Works and a Runnable Example

  What image recognition is, how classification differs from detection, and two runnable PyTorch and Keras examples with output explained.…

• Image Segmentation: Methods, Tools and How to Evaluate It

  Image segmentation methods in scikit-image, manual 3D segmentation in Slicer, and how to score a result with Dice and IoU.…

• Imitation Learning in Python: A Research Guide

  Imitation learning trains policies from expert demonstrations. Learn how to install the imitation library, execute behavioral cloning, and interpret l…

• Isaac Lab: Install, Isaac Sim Versions and a First Training Run

  Isaac Lab is NVIDIA's open-source robot-learning framework. Check Isaac Sim version fit, licences, install commands and a first training run.…

• Machine Translation: How It Works, Types and Research Use

  How machine translation works, its main types, when researchers can rely on it, and a verified step-by-step API translation request.…

• MS COCO: Dataset Guide, API Setup, and Metrics

  A guide to MS COCO: downloading annotations, loading data with pycocotools, interpreting JSON schemas, and computing 12 standard detection evaluation…

• Physics-Informed Neural Networks (PINNs): DeepXDE Tutorial

  Run a physics-informed neural network (PINN) with DeepXDE: install, backend choice, a 1D Poisson example and how to read the error.…

• Robot Learning: Software, Training, and Evaluation

  Robot learning trains policies for physical robots. Learn how to set up LeRobot in PyTorch, run imitation learning, evaluate metrics, and troubleshoot…

• Saliency Maps: PyTorch and Captum Research Guide

  Compute and visualize saliency maps for deep neural networks using PyTorch and Captum, including gradient aggregation, visualization, and troubleshoot…

• SHAP (SHapley Additive exPlanations): Computing and Reporting SHAP Values in Python

  SHAP in Python: install, compute SHAP values, read waterfall and beeswarm plots, and report the settings a paper needs.…

• Speech Recognition: How It Works and How to Run It (Whisper, Python)

  Speech recognition explained for researchers: how it differs from voice recognition, plus install and run commands for Whisper and the Python package.…

• Stable-Baselines3: Install, Train and Evaluate RL Agents

  Install Stable-Baselines3 from the official docs, train and evaluate an RL agent, and record the versions a methods section needs.…

• Surrogate Models: SMT Python Workflow Guide

  A guide to surrogate models in Python using SMT: Latin Hypercube design of experiments, Kriging model training, variance estimation, and troubleshooti…

• Train Test Split in Python: scikit-learn train_test_split Explained

  Train test split in Python with scikit-learn: the train_test_split signature, how test_size, shuffle and stratify interact, and split-first rules.…

• URDF: Robot Modeling and Validation Guide

  URDF defines robot kinematics, dynamics, and visual meshes for ROS. Learn how to write links, joints, and validate syntax using standard diagnostic to…

• WACV: IEEE/CVF Winter Conference on Applications of Computer Vision

  WACV 2027 explained: organizers, January 2027 dates, two submission rounds, and how the Applications and Algorithms tracks are reviewed.…

Scholarly Publishing

• AAAI: Artificial Intelligence Society and Conference Guide

  An operational guide to AAAI covering society governance, AAAI conference submission workflows on OpenReview, peer review tiers, and tier comparisons.…

• AISTATS Conference: AI and Statistics Guide

  A definitive research guide to AISTATS, documenting its CORE Rank A status, OpenReview submission guidelines, PMLR proceedings, and acceptance rates.…

• alphaXiv: Interactive arXiv Research Platform Guide

  An operational guide to alphaXiv covering its interactive arXiv preprint layer, inline annotations, verified author Q&A, and AI research tools.…

• ArXiv Sanity

  ArXiv Sanity and its successor ArXiv Sanity Lite provide machine learning researchers with custom preprint filtering, TF-IDF feature extraction, and S…

• Hugging Face Papers: Guide to AI Research on HF

  A comprehensive guide to Hugging Face Papers, detailing its arXiv indexing mechanisms, model and dataset linkages, author claiming, and platform workf…

• ICLR: Machine Learning Research Publishing Guide

  An operational guide to ICLR covering OpenReview submission workflows, reciprocal reviewer obligations, rebuttal strategies, and AI conference compari…

• IJCAI: International Joint Conference on AI

  IJCAI is the premier international artificial intelligence research conference, featuring rigorous double-blind peer review, CORE A* ranking, and the…

• MICCAI: Conference, Society, Proceedings and Open Access

  MICCAI explained: what it stands for, 2026 Strasbourg and 2027 Auckland details, 2026 acceptance rate, and where proceedings and open-access papers li…

• MINIMAR: The Reporting Checklist for Clinical AI and Machine Learning Models

  MINIMAR (MINimum Information for Medical AI Reporting) is a 2020 JAMIA checklist covering the minimum information a clinical AI/ML model's reporting n…

• NeurIPS Acceptance Rate: Trends and AI Benchmarks

  Analyze verified NeurIPS acceptance rates from official fact sheets, tracking submission surges, reviewer scaling, and comparisons with ICML, ICLR, an…

• Papers with Code: ML Benchmarks and Code Guide

  A research guide to Papers with Code, documenting its platform history, paper-to-code dataset dumps, SOTA benchmarks, and modern alternatives.…

• PMLR: Proceedings of Machine Learning Research

  Proceedings of Machine Learning Research (PMLR) is an open-access series publishing peer-reviewed proceedings for flagship AI conferences including IC…

• Stanford AI Index: Data, Metrics, and Research

  Explore the Stanford AI Index from Stanford HAI. Learn what it covers, how to access its open datasets, technical benchmarks, and research application…

• Transactions on Machine Learning Research Review

  Is Transactions on Machine Learning Research peer reviewed and Scopus indexed? Examine TMLR review criteria, OpenReview workflow, and verified journal…

• UCI Machine Learning Repository: What It Is, How to Search and Cite It

  What the UCI Machine Learning Repository is, how to filter and import its datasets, and how to cite and document them for reproducible research.…

Research Data Management

• Common Crawl: Open Web Data Corpus Guide

  Explore Common Crawl for AI and text mining. Learn CCBot crawl mechanics, WARC and WET formats, CDX index queries, AWS S3 access, and research licensi…

• Croissant: MLCommons' Metadata Format for ML-Ready Datasets

  Croissant is MLCommons' open JSON-LD metadata format, built on schema.org, for describing ML-ready datasets, their features, splits, and labels, so ML…

• ImageNet: Architecture, Access, and ML Benchmarks

  ImageNet guide for researchers: explore dataset scale, ILSVRC benchmark splits, institutional access terms, synset hierarchies, and privacy audits.…

• PhysioNet Credentialed Access: How to Get MIMIC and Other Restricted Data

  How PhysioNet's open, restricted, and credentialed access tiers work, and what the credentialing process (CITI training, identity review, and a per-da…

Research Methods & Statistics

• Artificial Intelligence vs Data Science: What Is the Difference?Comparison

  AI builds systems that perceive, reason, learn and act; data science extracts insight from data. They overlap in machine learning but differ in goals,…

• Machine Learning vs Artificial Intelligence: What Is the Difference?Comparison

  Machine learning is a subset of artificial intelligence: AI is the broad field, ML is the data-driven approach inside it. Side-by-side comparison of s…

• What Is Artificial Intelligence? Research Areas, Funding, and Career Paths

  A researcher-focused guide to artificial intelligence: what it studies, its major sub-fields, the real federal funding landscape (NSF, NIH, DARPA, DOE…

• What Is Machine Learning? Research Methods, Pitfalls, and Funding

  What machine learning is, its subfields, how to evaluate models without data leakage, who funds the research (NSF), and the key conferences and journa…

Clinical Research Administration

• AI Reporting Guidelines: CONSORT-AI, TRIPOD+AI, DECIDE-AI and the Rest

  A decision guide to the AI-specific reporting-guideline family: which of CONSORT-AI, SPIRIT-AI, TRIPOD+AI, DECIDE-AI and CLAIM applies to your study s…

• FDA AI Guidance: The Regulatory Framework for Artificial Intelligence in Drug Development and Medical Devices

  A guide to FDA's two parallel AI guidance tracks: CDRH's SaMD AI/ML and PCCP framework for medical devices, and CDER/CBER's draft guidance on AI in dr…

• FDA Predetermined Change Control Plans (PCCP) for Adaptive AI: What to Include

  How to draft an FDA Predetermined Change Control Plan (PCCP) for an AI-enabled device: the three required components (Description of Modifications, Mo…

Grants Management & Research Funding

• DARPA AI: Programs, Funding Thrusts & Guide

  A research administrator guide to DARPA AI funding, covering the AI Next campaign, AIE fast-track awards, BAA submission procedures, and compliance st…

• Schmidt Sciences: AI Safety, Trustworthy AI, and 2026 Funding Programs

  Schmidt Sciences is the 2024 successor to Schmidt Futures and funds AI2050 and the Science of Trustworthy AI RFP line — distinct from the Schmidt Scie…

Laboratory Compliance & Quality

• Software as a Medical Device (SaMD): Definition, Classification, and Regulatory Pathways

  Before any FDA pathway question comes a prior one: is the software a device at all? The 21st Century Cures Act section 3060 carve-outs — clinical deci…

Research Information & Identifiers

• Persistent Identifiers and Citation for AI Models and Training Datasets

  How to mint a persistent identifier for a trained AI model or training dataset, what a model card needs to document to make that identifier meaningful…

Scholarly Writing

• ICML Position Paper: Structure and Guidelines

  A practical guide to drafting an ICML position paper, covering mandatory title conventions, bold thesis requirements, alternative views, and review ru…

Key terms

• AI bias
• AI fairness
• AI safety case
• Bias audit (model)
• Compute (FLOPs estimate)
• Datasheet for datasets
• EU AI Act Article 53 (General-Purpose AI Model Obligations)
• Foundation model
• Frontier model
• Generative AI
• Hugging Face Hub (concept)
• Large language model (LLM)
• MLCommons benchmark
• Model audit
• Model card
• Model checkpoint
• Model evaluation suite
• Model fine-tune lineage
• Model lineage
• Model versioning
• Model weight licence
• National AI Research Resource (NAIRR)
• Neural Computation Journal Impact Factor (JIF)
• NIST AI RMF (Risk Management Framework)