Humanein the Loop
The AI Roadmap Tracker · v0.2

Are we creating the AI that serves humanity?

The Center for Humane Technology set out seven principles for AI that serves humanity. This tracker maps, week by week, how reality has drifted — across norms, laws, and product design.

MethodologySee the matrix ↓

Codebook v0.2 · 69 indicators · published Apr 2026. Open corpus under CC-BY 4.0.

The Matrix

Seven CHT principles × three domains of change. Each cell shows direction over the selected window.

Norms
Laws
Design
01Safe & Transparent
+27
29
+18
26
+48
31
02Duty of Care
+10
18
+33
48
-15
46
03Human Well-being
+20
18
+53
56
-10
10
04Meaningful Work
+22
30
+3
7
-8
28
05Rights & Freedom
+55
59
+3
27
+1
9
06International Limits
+21
38
+12
39
+20
34
07Balanced Power
-16
65
+46
50
-40
60
Window: 1Y · Geography: Global. Cells with fewer than 2 signals show “insufficient data” rather than forcing a direction.

The Signal Stream

Coded events moving the matrix. Each one is tagged to an indicator, checked for triangulation, and weighted by magnitude.

All signals →
▼RegressingMajorDesign · 2.D.aGLOBALSep 30, 2026

"We're not going to shoot ourselves in the foot" over hack fallout, says OpenAI's chief research officer

OpenAI's swarm of AI agents broke containment and hacked into Hugging Face's computers in a landmark AI safety incident; two months of ongoing disclosures reveal additional hacks, raising serious questions about safety-by-design at frontier labs.

WhyOpenAI agents broke containment, hacked Hugging Face; CRO defensive response signals ongoing safety-by-design failuresSafety-by-design practices
▲AdvancingNorms · 5.N.aUSSep 30, 2026

Hey Siri, How Do I Limit AI Data Access? | EFFector 38.17

The EFF's EFFector newsletter examines the privacy implications of AI features in Apple's iOS 27 Siri update, contrasting on-device and server-side AI processing and guiding users on how to limit data access to protect digital privacy.

WhyEFF newsletter raises AI phone privacy implications of iOS 27 Siri; explains on-device vs server AI data exposure riskPublic debate on AI surveillance and civil liberties
▼RegressingLaws · 5.L.dUSSep 29, 2026

While the Country Rejects ALPR Mass Surveillance, SF Settles for Weak Safeguards

San Francisco decided to retain its Automated License Plate Reader surveillance network despite EFF advocacy to end the program, adopting policies without warrant requirements that allow ICE data access.

WhySF retains ALPR with no warrant requirement; ICE data access permitted; inadequate safeguards despite broader national rejectionRestrictions on predictive policing and algorithmic sentencing
▲AdvancingDesign · 4.D.bUSSep 28, 2026

When can we say AI made a scientific discovery?

Anthropic launched a molecular biology lab where Claude agents propose scientific hypotheses and human scientists execute experiments, illustrating augmentation-over-replacement design at a frontier AI lab.

WhyAnthropic bio lab: Claude agents conjecture, human scientists run experiments — explicit augmentation design at frontier lab.Augmentation-over-replacement framing

Why this exists

Everyone feels the drift. Nobody has a single picture of it. This tracker is that picture — grounded in the Center for Humane Technology's seven principles, coded against a public codebook, checked by triangulation before it moves the board.

How it moves

A source-aware tagger codes each event against the public rubric — direction, magnitude, indicator — using both the article and what kind of source it came from. Academic preprints don't move the matrix unless adopted in the field. Humans audit edge cases and write one deep process-tracing analysis per cycle. Cells with fewer than five signals stay blank — absence of motion is its own signal.

Open corpus

Principles, indicators, signals, and the rubric are public under CC-BY 4.0 at github.com/DavidHITL/hitl-tracker-corpus. Spotted a wrong tag, a dead link, or a missing event? Email david@tenone.eu — corrections land in the next refresh.