Humanein the Loop
← The matrix
04

AI should not automate away meaningful work and human dignity

Current path

AI is framed as a replacement for human work without worker voice in deployment decisions. Algorithmic management erodes dignity. Automated decision systems operate without explanation or contestability.

Better future

AI augments meaningful work and preserves human dignity. Workers have voice in deployment decisions. Consequential automated decisions are explainable and contestable.

Drift across the three domains

Norms

Advancing3 signals
CHT recommends
  • Include worker voice in AI deployment decisions as a professional standard.
  • Reject inevitabilist replacement framings in favor of augmentation.
Indicators we track
  • 4.N.aPublic discourse on AI and labor
  • 4.N.bWorker voice in AI deployment decisions
  • 4.N.cPublic skepticism toward AI replacement rhetoric

Laws

Advancing5 signals
CHT recommends
  • Fund transition support and retraining for workers displaced by AI.
  • Regulate algorithmic management of workers.
  • Require notification and consultation before consequential AI deployment in workplaces.
  • Codify rights to explanation and human review of automated decisions.
Indicators we track
  • 4.L.aWorker displacement protections and transition funding
  • 4.L.bAlgorithmic management regulations
  • 4.L.cNotification requirements before AI deployment in workplace
  • 4.L.dAutomated decision-making rights

Design

·
Insufficient data1 signal
CHT recommends
  • Keep humans in authoritative roles for high-stakes decisions.
  • Position products as augmenting rather than replacing human work.
  • Provide transparent attribution of AI-generated work.
Indicators we track
  • 4.D.aHuman-in-the-loop for consequential decisions
  • 4.D.bAugmentation-over-replacement framing
  • 4.D.cTransparent attribution of AI-generated work

Recent signals

AdvancingLaws · 4.L.aUSMar 17, 2026

Minnesota introduces HF 4369 (Safeguarding Human Intelligence Act) with 90-day AI displacement notice

HF 4369 introduced 17 March 2026 requires employers to notify workers 90 days before AI-driven layoffs and fund retraining; first US state-level AI displacement notification bill. SF 4576 companion filed in Senate.

WhyConcrete state-level protection template; shifts displacement from externality to regulated transitionWorker displacement protections and transition funding
AdvancingLaws · 4.L.aUSDec 3, 2025

AI Workforce PREPARE Act introduced in US Senate

S.3339 introduced 3 Dec 2025 by Sen. Jim Banks (R-IN) with cosponsors Hassan (D-NH), Hickenlooper (D-CO), and Husted (R-OH); establishes federal AI workforce transition fund and retraining grants. Bipartisan cosponsorship signals emerging consensus.

WhyFederal-level displacement funding mechanism; first bipartisan AI transition bill with serious cosponsors across party linesWorker displacement protections and transition funding
AdvancingMajorNorms · 4.N.bEUROPENov 11, 2025

EU Parliament EMPL committee adopts algorithmic management directive call

Employment and Social Affairs Committee MEPs voted 11 Nov 2025 (41-6-4) to urge Commission to propose binding rules on algorithmic management across all sectors, extending Platform Work Directive protections to traditional employment. Full plenary endorsed the resolution 17 Dec 2025.

WhyWorker voice institutionalized at EU level; binding framework for surveillance/monitoring limits advancesWorker voice in AI deployment decisions
AdvancingDesign · 4.D.cUSOct 13, 2025

California AB 853 delays SB 942 AI Transparency Act effective date to August 2026

SB 942 (signed 2024) would have required generative AI providers serving >1M Californians to offer free AI-detection tools and mandatory provenance disclosures from 1 Jan 2026. AB 853, signed 13 Oct 2025, pushed the effective date to 2 Aug 2026 citing implementation readiness. Mixed signal: law still advances, but compliance delayed under industry pressure.

WhyFirst US state AI-content provenance regime still on track, but eight-month delay reveals industry pushback powerTransparent attribution of AI-generated work