# 07 — Sources & Bibliography

**Research window:** 21 October 2026
**Method:** Primary guidelines were read directly; secondary sources were used to understand the 2026 "AI slop" discourse and portfolio expectations. Some pages (Awwwards, ACM DL) blocked automated access; their claims were not used unless corroborated elsewhere.

---

## Primary — company design guidance

1. **Apple — Human Interface Guidelines: Design principles**
   <https://developer.apple.com/design/human-interface-guidelines/design-principles>
   Used for: Purpose, Agency, Responsibility, Familiarity, Flexibility, Simplicity, Craft, Delight.

2. **Apple — Human Interface Guidelines: Generative AI**
   <https://developer.apple.com/design/human-interface-guidelines/generative-ai>
   Used for: responsible AI, user control, transparency, clear expectations, non-AI fallback.

3. **Apple — Human Interface Guidelines: Machine Learning**
   <https://developer.apple.com/design/human-interface-guidelines/machine-learning>
   Used for: critical vs complementary ML, private vs public data, error handling.

4. **Apple — Liquid Glass technology overview**
   <https://developer.apple.com/documentation/technologyoverviews/liquid-glass>
   Used as context for 2025–26 material trends; not adopted as a visual style.

5. **Google Design — Expressive Design: Google's UX Research**
   <https://design.google/library/expressive-material-design-google-research>
   Used for: 46 studies, 18,000+ participants, 87% preference among 18–24, colour/shape/size/motion/containment, accessibility.

6. **Google — Material 3 Expressive**
   <https://m3.material.io/blog/building-with-m3-expressive>
   Used as the official system reference (page is JS-rendered; content corroborated via Google Design article).

7. **OpenAI — Design Guidelines / Brand**
   <https://openai.com/brand/>
   Used for: OpenAI Sans, perfect-circle wordmark, Blossom geometry (circles + right angles), open space, no effects/textures.

8. **Anthropic — design tokens (third-party documentation, updated 6 Oct 2026)**
   <https://www.designmd.co/d/anthropic>
   <https://styles.refero.design/style/d469cba4-c448-4a43-a033-883f8bfcdc42>
   Used for: `#141413` ink, `#FAF9F5` paper, Anthropic Sans, 2px spacing increments, editorial warmth.
   *Note: Anthropic does not publish a complete public token file; these are community-extracted values, used as directional evidence, not copied verbatim.*

9. **Microsoft — Guidelines for Human-AI Interaction (HAX Toolkit)**
   <https://www.microsoft.com/en-us/haxtoolkit/ai-guidelines/>
   Origin paper: <https://www.microsoft.com/en-us/research/publication/guidelines-for-human-ai-interaction/>
   Used for: 18 evidence-based guidelines across initial interaction, during interaction, errors, and over time.

---

## Secondary — 2026 UX & portfolio research

10. **Nielsen Norman Group — State of UX 2026: Design Deeper to Differentiate** (16 Jan 2026)
    <https://www.nngroup.com/articles/state-of-ux-2026/>
    Used for: AI hype → fatigue; "UI is still important, but it'll gradually become less of a differentiator"; trust requires transparency, control, consistency and support.

11. **RemoteWorks — Portfolio Design Trends for 2026**
    <https://remoteworks.pro/blog/portfolio-design-trends-2026>
    Used for: minimalism winning; dark mode default for technical audiences; case studies over galleries; mobile-first; 3–4 strong case studies beat ten shallow entries.

12. **Figma — Top Web Design Trends for 2026**
    <https://www.figma.com/resource-library/web-design-trends/>
    Used for: retrofuturism, collage, neo-minimalism, expressive colour, immersive 3D — and to decide which trends to reject.

13. **Behance — Design Trends 2026**
    <https://www.behance.net/gallery/239027109/Design-Trends-2026>
    Used for: emotion-led humanised design, visible grids, asymmetry, intentional imperfection, accessible vibrancy.

14. **925Studios — AI Slop Web Design: Complete Guide** (19 Mar 2026)
    <https://www.925studios.co/blog/ai-slop-web-design-guide>
    Used for: definition of AI slop, distributional convergence, default Inter/purple gradients/rounded cards, 38% visitor loss claim, human content 94.12% claim (secondary citations treated cautiously).

15. **Gi Pyeong Lee — Why Do All AI-Generated Websites Look the Same? (Hallmark)** (27 Jul 2026)
    <https://gipyeong-lee.github.io/2026/07/27/Hallmark-Anti-AI-Slop-Design-Skill-for-Claude-Code-Cursor-and-Codex.en/>
    Used for: Hallmark's refuse → apply → test model and the 57–65 point slop test.

16. **NxtFruit — AI Slop Design** (8 Aug 2026)
    <https://nxtfruit.com/2026/08/08/ai-slop-design/>
    Used for: "technically fine but visually interchangeable"; homogenisation. *(Page timed out during scraping; only the search excerpt was used.)*

17. **DesignMD — Taste Skill / anti-slop layout rules**
    <https://www.designmd.co/taste-skill>
    Used as a cross-check for layout, typography and spacing rules.

---

## Academic / evidence-based (context)

18. **Guidelines for Human-AI Interaction** — Amershi et al., CHI 2019 (18 guidelines; basis of Microsoft HAX).
    <https://www.microsoft.com/en-us/research/publication/guidelines-for-human-ai-interaction/>

19. **User Experience Design Professionals' Perceptions of Generative AI** — CHI 2024.
    <https://arxiv.org/pdf/2309.15237>
    Used as background on designers' trust and control concerns.

20. **How AI-generated images affect human preference and trust in websites** — ACM, Apr 2026.
    <https://dl.acm.org/doi/10.1145/3772363.3798849>
    *(Abstract blocked by Cloudflare; not quoted. Listed for completeness and future reading.)*

---

## Local evidence

21. **Project inventory** — `documentation/00-project-inventory.md`
    Generated from the local `C:\Users\mahes\Desktop\all projects` folders: READMEs, manifests, build artifacts and git remotes.

22. **Existing site** — `index.html`, `styles.css`, `script.js` (pre-redesign), plus `project_analysis.md`.
    Used for project history, descriptions and facts to verify.

23. **GitHub public profile** — <https://github.com/maheshwarkibehan-hub>
    Public API read on 21 Oct 2026: name, bio (`maheshwar.xyz`), 9 public repos.

---

## Limitations

- Firecrawl credits were over the plan limit; bulk reading was done with crawl4ai and targeted Firecrawl queries. All quotes were verified against the returned page text.
- Awwwards and ACM DL blocked automated access. No claims depend on them.
- Anthropic's tokens are community-extracted. They informed the direction (warm paper/ink, editorial type), not exact values.
- Statistics cited by agencies (38%, 94.12%) are secondary; they are used as directional evidence, not as proof.
- Design trends are dated by nature. The principles in docs 01–05 are intended to outlive the trends.

---

## Citation style

In the documentation and code comments, refer to sources as `[S01]`…`[S23]` using the numbers above. No source text is reproduced beyond short quotations for analysis.
