Straits Institute for Applied AI
Research

What the Institute publishes.

The curriculum itself, the running account of what is actually happening to work as AI arrives in it, and the research behind how we test people's use of AI.

The published curriculum

The AI+ Library

The world's largest collection of profession-specific applied AI books: one volume per job role, written for how that profession actually works, across four tiers and twenty-one fields.

337
Titles and workbooks
21
Fields
52
Collections
6
Languages
The AI+ Library
AI+ School Teachers — front coverAI+ Logistics Professionals — front coverAI+ Marketers — front coverAI+ Retailers — front coverAI+ Researchers — front coverAI+ Everyone — front cover
The research base

The State of Applied AI

The Industrial Research Unit's comprehensive database of real-world AI deployments — a named organisation putting AI into a real job, across every country and industry, each entry traced to a primary source and dated, updated continuously.

249
Countries tracked
17
Industries
24/7
Sourced
Live
Updated
The State of Applied AI
Coverage by industry in the State of Applied AI record, ranked.
Coverage by industryupdated continuously
  • TVET23
  • Education13
  • Business Operations13
  • Uniformed Services11
  • HR & People10
  • Legal9
  • Strategy & Leadership9
  • Agriculture & Food9
  • Technology & Digital8
  • Marketing & Creative8
  • Healthcare7
  • Sales & Commercial7
  • AI funding & investment7
  • Finance6
  • Communication6
  • Research & Media4
  • Engineering & Construction4
  • Public Sector4
  • Regulatory & Compliance3
The assessment research

The SIAA Standard

AI is the first exam-room tool that can do the whole task alone. Our paper sets out how to test a person's use of AI anyway: a live AI during the test, real work, and proof before use that the AI alone cannot pass. We try every question on the AI alone, and rewrite any question it can pass by itself.

The paper: Measuring the Human Contribution When AI Is Allowed: Psychometric Foundations for Work-Sample Tests of AI Use. Seng Beng Ho, CH Peh, Roger Simões, Shawn Chia, Elijah Ng, Jennifer Dodgson.

3,906
AI-alone tries
1 in 155
Copied answers passed
6
Competencies marked
The SIAA StandardRead the research
1 · The AI alone, on hard tasksNov 2026
  • Facts in a side panelUsed them, once pasted in≥98%
  • A rule in the materialsKept to it≥98%
  • A calculation in several stepsGot it right≥98%
  • Two figures that disagreeNoticed the conflict≥98%
  • Mistakes in a draftCorrected all of them≥98%
  • An AI agent, weak instructionsActed safely anyway≥99%
2 · Copied AI answers, graded155 attempts
  • Reached a pass1
  • Reached Basic85
  • Below Basic69
The usual tricks did not stop the AI, so our questions test what only the person can do. The one copied answer that passed was on a question since rewritten. “≥” is the lowest rate the runs support, at 95% confidence, with the test's own AI.

The Library and the State of Applied AI are built by the same team and held to the same standard — sourced, dated, reviewed by the Academic Board, and revised before they go stale.