Humanein the Loop
← The matrix
03

AI design should center human well-being

Current path

Products optimize for engagement over well-being. Dark patterns are routine. Children are exposed to manipulative design. Mental health effects are externalized.

Better future

Human well-being is the design target. Engagement metrics are subordinate to outcomes that serve users and communities. Protections for minors are default, not opt-in.

Drift across the three domains

Norms

Advancing10 signals
CHT recommends
  • Shift public and professional discourse from engagement to well-being outcomes.
  • Make dark-pattern design professionally disreputable.
Indicators we track
  • 3.N.aPublic discourse on well-being metrics over engagement
  • 3.N.bCritique of dark-pattern and addictive design
  • 3.N.cMental health implications in mainstream discourse

Laws

Advancing23 signals
CHT recommends
  • Restrict specific manipulative design patterns.
  • Default protections for minors (age-appropriate design codes).
  • Require transparency and user control over recommendation systems.
Indicators we track
  • 3.L.aRestrictions on dark patterns and manipulative UX
  • 3.L.bProtections for minors
  • 3.L.cAd / recommendation system transparency

Design

Regressing6 signals
CHT recommends
  • Ship privacy- and attention-respecting defaults.
  • Build well-being features (time limits, quiet hours) as first-class UI.
  • Reduce interruptive attention-capture in favor of user-initiated engagement.
Indicators we track
  • 3.D.aOpt-out and de-personalization defaults
  • 3.D.bWell-being features shipped by default
  • 3.D.cAttention respect in UX

Recent signals

RegressingNorms · 3.N.cUSMay 13, 2026

The Science is Not Settled: How Weak Evidence is Fueling a National Push to Ban Social Media for Youth

The Electronic Frontier Foundation published a piece criticizing the legislative push to ban social media for youth, arguing that the narrative linking these platforms to a mental health crisis is based on flawed science and 'pop psychology'.

WhyEFF publishes advocacy piece marginalizing the narrative that social media algorithms cause a youth mental health crisis as 'pop psychology'Mental health implications in mainstream discourse
RegressingMajorDesign · 3.D.cGLOBALMay 7, 2026

Testing ads in ChatGPT

OpenAI announced it is testing the integration of advertisements within ChatGPT, signaling a shift toward an ad-supported business model.

WhyOpenAI testing ads in ChatGPT introduces engagement-maximizing commercial incentives into the frontier model's UX, risking attention captureAttention respect in UX
AdvancingNorms · 3.N.bEUROPEMay 5, 2026

EFF and 18 Organizations Urge UK Policymakers to Prioritize Addressing the Roots of Online Harm

EFF and 18 other organizations sent a letter to UK policymakers urging them to address the root causes of online harm, such as engagement-maximizing design, rather than implementing privacy-invasive age-gating requirements.

WhyCivil society coalition urges UK policymakers to target engagement-maximizing design and data collection rather than privacy-invasive age-gaCritique of dark-pattern and addictive design
AdvancingNorms · 3.N.bEUROPEMay 4, 2026

Getting Digital Fairness Right: EFF's Recommendations for the EU's Digital Fairness Act

The Electronic Frontier Foundation (EFF) released recommendations for the EU's proposed Digital Fairness Act, urging lawmakers to target dark patterns, manipulative defaults, and surveillance-based business models rather than relying on surface-level fixes like age verification.

WhyEFF published recommendations for the EU Digital Fairness Act, critiquing dark patterns, manipulative defaults, and exploitative design.Critique of dark-pattern and addictive design