Artificial Intelligence Index Report 2026

Stanford Institute for Human-Centered AI (AI Index Steering Committee) · 2026 · Stanford HAI evidence medium priority coded

Main argument

Annual empirical index: technical performance (reasoning/coding benchmarks approaching saturation), economy/adoption (88% organizational adoption; consumer diffusion), policy/regulatory attention, responsible AI (Foundation Model Transparency Index average FELL to 40 from 58 - developers disclosing less as capability rises), public opinion chapters - with the headline that the capability-governance gap is WIDENING.

Why it matters here

The empirical almanac: adoption (88% organizational), capability jumps (SWE-bench 60%->~100% in a year), transparency DECLINE (Foundation Model Transparency Index 58->40), responsible-AI and public-opinion chapters. Use for empirical framing claims - especially the transparency decline, which cuts directly against every explanation-based responsibility regime coded in the library.

Reading notes

Reference-report registration (38MB; responsible-AI + public-opinion chapters are the relevant slices). Key 2026 numbers logged from the report's own summary: near-saturation coding benchmarks, 88% org adoption, 4/5 students using genAI, FMTI average down to 40 from 58, 'widening gap between what AI can do and how prepared we are to manage it'.

Stanford HAI. (2026). Artificial Intelligence Index Report 2026. Stanford Institute for Human-Centered AI.

Close reading — 1 coded units

#1 · pp. 1 · evidence

“[2026 Index headlines: SWE-bench Verified from 60% to near 100% in a year; 88% organizational adoption; Foundation Model Transparency Index average dropped to 40 from 58; 'a widening gap between what AI can do and how prepared we are to manage it'.]”

RL-EPISTEMIC

Synthesis-matrix row

supports T6-RESPONSIBILITY-UNALLOCATED

transparency decline erodes the epistemic basis for any allocation

Memos (1)

thesis-link · unit #1

The FMTI decline (58->40) is the single most useful number in the report for the dissertation: every explanation-based responsibility mechanism in the coded library (Kästner's MI regime, S&K's justification criterion, Huang's knowledge-condition interface, J&N's contestability) PRESUPPOSES disclosure - and the measured trend is the opposite. This grounds the governance chapter's case that voluntary transparency has failed and responsibility-allocation rules must be mandatory (EU AI Act art. 86 right-to-explanation; H&D's liability clarification). Cite alongside the agentic-capability numbers to show both blades of the scissors: capability up, epistemic access down.