L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Module 5 · Lesson 1

How AI Reads and Generates Type

From pixels to embeddings — what language models and vision models actually see when they look at text
Does AI understand typography, or just imitate it?

In 2023, Adobe began shipping Firefly with a text-effects feature that let users type a word and receive it rendered in a stylised texture — fire, water, leaves. The feature went viral. What most designers didn't notice was that Firefly was not choosing a typeface at all: it was generating pixels shaped like letterforms, trained on millions of licensed font specimens. The model had learned the statistical shape of letters without ever holding a type rule.

Two Very Different Problems

When we say "AI and typography" we are actually describing at least two distinct technical problems that are often conflated. The first is type recognition — can a model identify which typeface is present in an image, or extract text from a layout? The second is type generation — can a model produce new letterforms, font files, or typographic layouts? Each relies on a different architecture.

Type recognition is largely a computer vision task. Models like Google's Vision API or Adobe's Sensei Font Finder use convolutional neural networks (CNNs) trained on databases of digitised font specimens to match glyph shapes. WhatFont and services like What the Font (MyFonts) use similar approaches. These tools are genuinely useful but have known failure modes: they struggle with hand lettering, distorted text, and typefaces released after their training cutoff.

Type generation is more complex. Diffusion models such as Stable Diffusion and Midjourney learn to generate pixels that look like text, but they do not understand letterforms at a structural level. This is why, until late 2023, AI image generators famously failed to produce legible text — they were hallucinating plausible-looking glyphs rather than rendering actual characters. The breakthrough came from conditioning approaches: DALL-E 3 (October 2023) was the first widely-deployed consumer model to produce reliably legible short text strings, achieved by training with dense captioning that explicitly described text content.

Real Event

In October 2023, OpenAI's release notes for DALL-E 3 specifically cited improved text rendering as a flagship capability. Independent testing by design publications including Creative Bloq and The Verge confirmed that short words (under ~8 characters) rendered legibly in most generations — a marked departure from earlier versions. Longer strings and unusual fonts remained unreliable.

Tokenisation and the Typography Gap

Large language models (LLMs) like GPT-4 and Claude process text as tokens — chunks of characters, typically 2–6 characters long. This means they have no inherent concept of a glyph's visual form. When you ask an LLM to "choose a font for a luxury brand," it is drawing on textual associations learned during training (reviews, design articles, brand guidelines published online) rather than any visual analysis. It knows that Didot appears in Vogue editorial contexts because that association appeared in its training corpus, not because it has looked at the letterforms.

This creates a practical implication for designers: LLMs are excellent at typographic reasoning about associations, hierarchy, and intent, but they cannot directly perceive whether a specific font pairing will work visually. Multimodal models (GPT-4V, Claude 3 Opus with vision) can analyse screenshots of type settings, but their visual perception of fine spacing and weight relationships is still less reliable than a trained human eye.

Token
A unit of text (word fragment, word, or punctuation) that an LLM processes. Not a glyph — no visual information attached.
Glyph
The specific visual representation of a character in a given typeface. Distinct from the abstract character (Unicode codepoint).
Embedding
A numerical vector representing a word or image patch. Vision models embed letterforms as spatial feature vectors.
Diffusion Model
An AI that generates images by iteratively denoising random pixels, guided by a text or image prompt.
Font Identification Tools in Practice

Several production-grade AI font-identification tools exist today. Adobe Fonts' visual search (powered by Sensei) lets designers upload an image and returns ranked font matches. MyFonts' WhatTheFont has been running since 2004 and was updated with deep learning in 2017, dramatically improving accuracy on clean printed specimens. Font Ninja and WhatFont browser extensions identify typefaces directly from live web pages by reading CSS font-family declarations — a different, non-AI approach that is simply more reliable for screen contexts.

The critical limitation for all vision-based font ID tools is kerning and OpenType features. A model might correctly identify that a headline is set in Garamond, but it cannot tell you whether the designer has applied optical kerning, activated contextual alternates, or set it with a non-standard tracking value. That interpretive layer remains entirely human.

Designer Takeaway

AI sees typography differently than you do. Use AI font tools for identification and association tasks. Use your own eyes — and your knowledge of type principles — to evaluate the visual result. The two capabilities are complementary, not redundant.

Quiz — Lesson 1

How AI Reads and Generates Type · 3 questions
1. DALL-E 3 (2023) improved text rendering primarily through which approach?
✓ Correct — Correct. OpenAI's release notes and contemporaneous reporting confirmed that dense, descriptive captioning — which explicitly named text appearing in training images — was the key change enabling DALL-E 3's improved text legibility.
Not quite. The improvement was specifically about dense captioning that named text content in training images, not a different architecture or a post-processing OCR step.
2. When an LLM recommends "use Didot for luxury branding," it is drawing on:
✓ Correct — Correct. LLMs process text tokens, not glyphs. Their typographic "knowledge" is entirely associative — patterns learned from design writing, not visual perception.
Remember: LLMs tokenise text and have no inherent glyph-rendering capability. Their font recommendations come from textual associations in the training corpus.
3. Which tool identifies typefaces from live web pages by reading CSS declarations — NOT using AI vision?
✓ Correct — Correct. WhatFont reads font-family CSS properties directly from the DOM — no image analysis involved. It's faster and more accurate for live pages precisely because it sidesteps the visual recognition problem.
The WhatFont extension works differently from image-based tools: it reads CSS font-family declarations directly, making it more reliable for web type identification.

Lab 1 — AI Type Perception

Explore how AI models understand (and misunderstand) letterforms and font identity

Your Challenge

Use the AI assistant below to explore the gap between how humans and AI models perceive typography. Ask it to recommend typefaces, explain why certain fonts carry specific associations, or describe what an LLM actually "sees" when processing typographic instructions. Push on its limits.

Starter prompt: "I'm designing a packaging label for an artisan gin. Walk me through how you'd reason about typeface selection — and be honest about whether you're making visual or associative judgments."
Type Perception Lab
AI for Graphic Design · M5 L1
Module 5 · Lesson 2

AI-Assisted Font Pairing

Using AI tools to shortlist, test, and refine type combinations — without ceding creative judgment
Can AI make better font pairings than an experienced designer? Under what conditions?

In 2021, Google Fonts launched its AI-powered font pairing suggestions — when a designer selects a display typeface in the Google Fonts interface, the tool proposes a ranked list of complementary body type options. The recommendations are generated by a model trained on co-occurrence data from millions of real websites that used Google Fonts. Typefaces that appeared together frequently on well-rated design projects were weighted positively. The result is a system that is pragmatically useful but culturally conservative — it recommends what has already worked, not what might be inventively different.

The Co-occurrence Model and Its Bias

Most AI font pairing tools work on some variation of co-occurrence analysis or collaborative filtering — the same techniques that power Netflix recommendations or Spotify's Discover Weekly. The model asks: what fonts have designers previously put together, in what contexts, and with what outcomes?

This approach has real strengths. It encodes the collective tacit knowledge of thousands of working designers. If you are unfamiliar with a typeface and need a safe pairing quickly, these tools genuinely reduce error rates. Fontjoy, which uses a neural network to generate pairings by measuring contrast along axes like x-height, weight, and classification, allows designers to lock one typeface and generate complementary suggestions — a genuinely useful workflow accelerator.

The limitation is systematic bias toward the mainstream. A model trained on popular websites will under-recommend expressive or experimental typefaces. It will suggest Playfair Display + Source Sans Pro far more often than a pairing that might be more appropriate for, say, a contemporary dance company's identity. The tool is calibrated for frequency, not fitness.

Real Tool: Fontjoy

Fontjoy (fontjoy.com), built by Jack Qiao in 2017 and updated incrementally since, uses a deep learning model to embed typefaces in a high-dimensional space based on visual features. Pairing works by finding typefaces with controlled contrast — similar enough to feel harmonious, different enough to establish hierarchy. Qiao documented the model architecture in a 2017 blog post, noting that the network was trained on ~1,800 Google Fonts and learned to encode stylistic features without explicit labels.

A Practical Workflow

The most productive approach treats AI pairing tools as a first-pass filter, not a final answer. A tested workflow used by several design agencies:

Brief the AI with contextAsk an LLM (ChatGPT, Claude, Gemini) to suggest typefaces given a detailed brand brief — industry, tone, audience, medium, competitors to differentiate from. The more specific the brief, the more useful the output.
Run the shortlist through Fontjoy or Google FontsTake the AI's 3–5 display suggestions and use a pairing tool to rapidly explore body type options. Lock the display face, generate 10–20 pairings, screenshot the interesting ones.
Apply to actual copyPaste the real headline and body copy into a test document. AI tools show generic placeholder text; real copy reveals fit problems immediately. Test at intended sizes and weights.
Interrogate edge cases with AIAsk an LLM to critique your shortlisted pairing — "what are the weaknesses of this combination for a B2B SaaS product?" This surfaces concerns you may have rationalised away.
Final decision is yoursNo AI tool has seen your specific client, their audience, or the competitive context you're working in. The final judgment requires your synthesis.
LLMs as Typographic Consultants

One underused application is prompt-based typographic critique. Because LLMs have ingested enormous amounts of design writing, they can reason about typographic conventions, historical associations, and potential misfires with surprising depth. Asking Claude or GPT-4 to "critique this typeface pairing for a healthcare brand" yields more nuanced feedback than most online resources, provided you give it enough context about the project.

The caveat: LLMs will sometimes confabulate — asserting facts about specific typeface histories or designer attributions that are inaccurate. Always verify any specific historical claim an LLM makes about a typeface against a primary source (the type foundry's specimen, Fonts In Use, or a reference like Bringhurst's The Elements of Typographic Style).

Key Principle

AI font pairing tools optimise for what has worked before. Your value as a designer lies partly in knowing when to follow that pattern and when to break it deliberately. Use these tools to compress the time spent on safe options, so you have more time to explore risky, distinctive ones.

Quiz — Lesson 2

AI-Assisted Font Pairing · 3 questions
1. Google Fonts' AI pairing suggestions are based primarily on:
✓ Correct — Correct. The pairing model was trained on co-occurrence patterns from millions of websites — typefaces that appeared together on well-rated projects were weighted positively.
Google Fonts pairing recommendations come from co-occurrence analysis — which fonts appeared together on real, successful websites — not from explicit rules or visual geometry alone.
2. Fontjoy generates pairings by finding typefaces with:
✓ Correct — Correct. Jack Qiao's documented approach for Fontjoy specifically targets controlled contrast — the model looks for complementary rather than identical or maximally different typefaces.
Fontjoy aims for controlled contrast, not sameness or maximum difference. The pairing needs enough similarity for visual harmony and enough difference for typographic hierarchy.
3. The primary limitation of AI font pairing tools for expressive or experimental design work is:
✓ Correct — Correct. Co-occurrence models recommend what has appeared most often, which systematically under-weights expressive or unconventional choices — even when those choices would be more appropriate.
The core issue is a frequency bias: tools trained on popular websites will keep recommending popular pairings, which is conservative by nature regardless of the specific design context.

Lab 2 — Font Pairing Consultant

Use AI to build and critique a typographic system for a real brief

Your Challenge

You're designing a typographic system for a new brand. Use the AI below as a typographic consultant. Give it a specific brief (industry, tone, medium, audience) and work through display, body, and UI typeface selection. Then ask it to critique the combination you've settled on.

Starter prompt: "I'm designing a typographic system for a contemporary architecture firm targeting high-net-worth residential clients. The primary medium is a printed portfolio and a website. Suggest a display and body typeface pairing and explain the rationale for each choice."
Font Pairing Consultant
AI for Graphic Design · M5 L2
Module 5 · Lesson 3

Variable Fonts and Generative Type

How AI is reshaping typeface design itself — from parametric axes to neural font synthesis
What happens when the typeface itself becomes a generative artefact?

In 2022, Google's Brain team published research on DiffVG-adjacent techniques for vector font generation. Separately, the type design community was already experimenting with neural interpolation within variable font design space — using machine learning to generate intermediate masters along axes that humans hadn't explicitly drawn. Type designer Christoph Haag and researchers at Zurich's ZHDK were among those exploring how ML could extrapolate credible weight instances beyond the range a designer had manually drawn.

Variable Fonts: A Foundation for AI Experimentation

Variable fonts, introduced in the OpenType 1.8 specification in 2016 and achieving widespread browser support by 2020, are a prerequisite for understanding how AI interacts with type at the design level. A variable font encodes a design space rather than a single fixed instance — the font file contains masters (typically at extremes of each axis) and interpolation logic that allows a single file to render any point within the defined space continuously.

Common axes include wght (weight), wdth (width), ital (italic), and opsz (optical size). Some typefaces define custom axes — e.g. GRAD (grade) in fonts like Roboto Flex, which allows weight adjustment without affecting metrics.

AI enters this picture in two ways. First, neural networks can be used to suggest axis values based on context — a responsive typesetting system might automatically adjust optical size and weight based on viewport, ambient light, or even reading distance (Samsung has filed patents in this direction). Second, and more experimentally, AI can be used to generate entirely new intermediate masters that were never explicitly drawn by a human type designer.

Real Research: FontDiffuser

In 2023, researchers from multiple Chinese universities published FontDiffuser, a diffusion-model-based system for few-shot Chinese font generation. Chinese typeface design is an extreme case — a complete typeface may require 20,000+ glyphs, making manual design enormously costly. FontDiffuser demonstrated that a diffusion model trained on a small number of glyphs could extrapolate credible glyph designs for unseen characters, maintaining stylistic consistency. The paper was accepted at AAAI 2024. This technique is directionally applicable to Latin fonts with far smaller required glyph sets.

Neural Font Synthesis: Current State

Several research projects and early commercial tools now demonstrate neural font generation. DeepFont (Adobe Research, 2015) was an early CNN-based font recognition and synthesis system — primarily recognition-focused but establishing foundational architecture. More recent work includes:

GANs for glyph generation — generative adversarial networks can produce new glyph variants within a defined style. Researchers at University College London published work in 2021 on GAN-based font interpolation that could create stylistically consistent new typefaces by blending existing ones. The results were visually coherent at display sizes but revealed spacing and hinting problems at text sizes.

Vector-native generation — most AI image generators produce rasterised output, which is unusable for professional font production. Research projects like DeepVecFont (2021) and VecFusion (2023) generate vector outlines directly, producing actual Bézier paths rather than pixels. These are closer to being useful for type designers, though the curve quality still requires significant manual cleanup.

Raster Font Generation
  • Output is pixels, not paths
  • Requires vectorisation post-process
  • Loses hinting and metric precision
  • Suitable for concept exploration only
Vector Font Generation
  • Output is Bézier paths (SVG / UFO)
  • Curve quality still needs manual review
  • Preserves metric information
  • Closer to production-ready pipeline
Practical Applications for Designers Today

The neural font synthesis research is largely pre-production, but practical tools are emerging. Prototypo (since discontinued in its original form) was an early parametric type design tool that let designers manipulate stem widths, apertures, and other structural parameters to create fonts without drawing individual glyphs — a manual forerunner to what AI is now automating.

For working designers, the most immediately applicable area is responsive typography powered by variable fonts and AI-driven context detection. CSS now exposes variable font axes via font-variation-settings, and JavaScript can adjust these axes dynamically based on viewport data. Pairing this with an AI system that interprets reading context is a near-future design problem that type designers and interface designers will need to solve together.

Watch This Space

Neural font generation is at the research-to-production transition point as of 2024–2025. Type designers who understand these tools will be positioned to direct AI-assisted workflows rather than be displaced by them. The craft knowledge needed to evaluate, correct, and direct AI-generated type — spacing, rhythm, optical corrections — is not being automated.

Quiz — Lesson 3

Variable Fonts and Generative Type · 3 questions
1. Variable fonts encode a _______ rather than a single fixed instance.
✓ Correct — Correct. A variable font file contains design space masters and interpolation data, allowing any point within that space to be rendered continuously — not a collection of separate instances.
Variable fonts encode a design space — masters at the extremes of each axis plus interpolation logic. This allows continuous variation across weight, width, optical size, and custom axes within a single file.
2. FontDiffuser (2023) was significant for type design primarily because it addressed:
✓ Correct — Correct. Chinese typefaces can require 20,000+ glyphs, making full manual design prohibitively expensive. FontDiffuser demonstrated few-shot generation — extrapolating credible new glyphs from a small number of examples.
FontDiffuser's significance was specifically about Chinese typeface design, where a complete typeface may require more than 20,000 glyphs — making AI-assisted glyph extrapolation economically transformative.
3. Compared to raster font generation, vector-native generation (e.g. VecFusion) is more useful for type production because:
✓ Correct — Correct. Vector output (Bézier paths in SVG or UFO format) preserves the structural and metric information that font production requires. Raster output must be vectorised, losing precision in the process.
The key advantage is vector paths (Bézier curves) rather than pixels. Professional font production requires outline data and metrics — rasterised output requires costly re-vectorisation that degrades quality.

Lab 3 — Variable Font Design Space

Explore variable font axes, AI-assisted parameter selection, and generative type concepts

Your Challenge

Use the AI below to explore variable font design spaces and understand how AI tools interact with parametric type. Ask about specific variable font axes for a project, discuss how neural interpolation could extend a type system, or work through a responsive typography scenario where variable font parameters adjust to context.

Starter prompt: "I'm building a responsive editorial website. Explain how I could use variable font axes — specifically wght, opsz, and a custom axis — to create a typographic system that adapts to different screen sizes and reading contexts. What role could AI play in adjusting these parameters dynamically?"
Variable Font Lab
AI for Graphic Design · M5 L3
Module 5 · Lesson 4

Typographic Hierarchy and AI Layout

How AI layout tools interpret, enforce, and sometimes break typographic hierarchy — and how to stay in control
When AI sets the type, who is responsible for the hierarchy?

In late 2023, Adobe InDesign shipped a generative AI layout feature (part of Firefly integration) that could auto-populate a template with content from a brief, automatically setting headline, deck, and body type at AI-determined sizes and weights. Early beta testers reported that the system consistently produced visually adequate but typographically flat layouts — the hierarchy existed, but was timid. Headlines were sized conservatively; the grid was applied mechanically. The issue wasn't incorrect type setting; it was that the AI had no understanding of editorial intent — what should feel urgent, what should feel considered, what should command the reader's eye from across the room.

What AI Measures vs. What Hierarchy Communicates

Typographic hierarchy is fundamentally about directing attention in a sequence that serves the reader's comprehension and the content's intent. It operates through contrast — in size, weight, colour, spacing, and position. AI layout systems can measure and apply contrast along these axes, but they optimise for legibility (can the text be read?) rather than rhetoric (does the arrangement argue for something?).

Systems like Canva's Magic Layout (2023), Adobe Express, and the various AI design generators (Looka, Wix ADI, Framer AI) all produce layouts with demonstrable typographic hierarchy. A headline is bigger than body copy. Captions are set smaller. But these are categorical decisions — class-based rather than contextual. The AI knows "headline: large" but cannot decide that this particular headline should be set at 120pt across a full bleed, because that would be strange and right for this specific piece.

Documented Limitation: Canva Magic Design

When Canva launched Magic Design (late 2023), design reviewers including the creative community at Dribbble and design critics at Eye magazine noted a consistent pattern: AI-generated layouts defaulted to centred type, moderate size contrast, and safe colour-type combinations. The layouts were functional but rarely distinctive. Canva's own design team acknowledged this in a 2024 blog post, framing it as intentional — the system is calibrated for non-designers who need "good enough" rapidly, not for designers seeking creative differentiation.

AI as a Type-Setting Rule Enforcer

Where AI adds clear value is in enforcing typographic rules at scale. A human designer can specify a typographic system — 4 type styles, specific size ratios, specific tracking and leading values — and AI tools can apply that system consistently across hundreds of pages or template variations. This is the premise of tools like Frontify's AI brand application features and Supernova's design token propagation.

More specifically, AI can catch typographic errors that humans miss under time pressure: widows and orphans (single words stranded at the end of a column or top of a page), inconsistent paragraph spacing, wrong em-dash versus en-dash usage, non-matching quote marks, and tracking inconsistencies between weights. Adobe InDesign's built-in story editor and some AI-augmented proofing plugins now flag these automatically.

Widow
A single short line or word at the end of a paragraph that sits alone at the top of a column or page. Disrupts rhythm.
Orphan
The first line of a paragraph stranded alone at the bottom of a column. Leaves the paragraph's opening without context.
Type Style
A named, reusable typographic specification — typeface, size, weight, leading, tracking, colour. The unit AI layout systems use.
Editorial Intent
The rhetorical purpose of a typographic decision — what the arrangement argues, prioritises, or makes the reader feel. Not measurable by AI.
Prompting for Typographic Hierarchy

When using generative AI tools to set type or create layouts, the quality of the typographic hierarchy is directly related to the specificity of your brief. Vague prompts produce generic hierarchy. Specific prompts that describe the emotional register, the reading sequence, and the dominant visual element produce more intentional results.

For example, a prompt like "create a magazine spread" yields centred, conservative type. A prompt like "create a magazine spread where the headline should feel like it's shouting — large, heavy, set left with aggressive tracking — followed by a quiet, condensed deck that provides context before the body copy begins" gives the AI enough directional information to approximate editorial intent, even if it cannot fully understand it.

The practical workflow: use AI to rapid-prototype hierarchy variations quickly, then apply your own editorial judgment to select and refine. Never accept the first AI-generated hierarchy as final. The AI is showing you possible arrangements; you are deciding which arrangement is right.

The Designer's Irreplaceable Role

AI layout systems are, at best, sophisticated automatic typesetters. They can apply a system. They cannot create one. The typographic system — the hierarchy of meaning, the intended reading sequence, the emotional register of every size and weight decision — must be authored by you. The more specific you are in defining that system, the more useful AI becomes in applying it consistently.

Quiz — Lesson 4

Typographic Hierarchy and AI Layout · 3 questions
1. AI layout systems like Canva's Magic Design optimise primarily for:
✓ Correct — Correct. These systems can confirm that a hierarchy exists (headline is larger than body), but they optimise for categorical legibility rather than the editorial argument a specific arrangement should make.
AI layout tools produce readable hierarchies but optimise for legibility rather than the rhetorical register or editorial intent that distinguishes good typographic design from merely functional typesetting.
2. According to Canva's own 2024 blog post, Magic Design's conservative typographic defaults are intentional because:
✓ Correct — Correct. Canva explicitly framed this as a deliberate design decision for their primary user base — non-designers who need a functional result quickly, not design professionals seeking creative differentiation.
Canva acknowledged this was intentional: the tool is built for a non-designer user base that needs "good enough" results fast. It is not designed for the creative differentiation that professional designers require.
3. When prompting AI tools to create typographic hierarchy, the most effective approach is to:
✓ Correct — Correct. Specificity about emotional register, intended reading sequence, and the dominant visual element gives AI layout tools enough directional information to produce hierarchy that approximates editorial intent.
Generic prompts yield generic hierarchy. Describing the emotional register, reading sequence, and visual dominance hierarchy in your prompt gives the AI the directional information it needs to produce something closer to intentional typographic design.

Lab 4 — Hierarchy Prompt Workshop

Practice writing typographic hierarchy briefs that give AI tools enough intent to work with

Your Challenge

The quality of AI-generated typographic hierarchy depends almost entirely on the specificity of your brief. In this lab, practice writing detailed typographic intent briefs. Describe a layout scenario and ask the AI to define the type hierarchy for it — then critique and refine the result. Focus on specifying emotional register, reading sequence, and visual dominance.

Starter prompt: "I'm designing a full-page newspaper advertisement for an emergency appeal — a humanitarian charity asking for donations after a flood disaster. Define the typographic hierarchy for this page: what should hit first, what emotional weight should each level carry, what size/weight/tracking relationships would you specify, and why? Be specific about the intent behind each decision."
Hierarchy Brief Workshop
AI for Graphic Design · M5 L4

Module 5 Test

Typography and AI · 15 questions · Pass mark 80%
1. What was the specific technical change that enabled DALL-E 3 to produce reliably legible text in generated images?
✓ Correct — Correct. OpenAI's own documentation confirmed dense captioning as the key change.
Dense captioning — training images with explicit descriptions of text content — was the documented change. Review Lesson 1.
2. When an LLM recommends a typeface, its knowledge is based on:
✓ Correct — Correct. LLMs process tokens, not glyphs — their typographic knowledge is purely associative.
LLMs have no visual glyph perception. Their recommendations come from textual associations. Review Lesson 1.
3. MyFonts' WhatTheFont updated its font identification system with deep learning in which year?
✓ Correct — Correct. WhatTheFont integrated deep learning in 2017, significantly improving accuracy on clean specimens.
WhatTheFont's deep learning update was in 2017. Review Lesson 1.
4. Google Fonts' AI pairing suggestions are systematically biased toward mainstream recommendations because they are trained on:
✓ Correct — Correct. The co-occurrence approach weights what has appeared most often, producing conservative, mainstream suggestions.
The bias comes from frequency-based co-occurrence training. Review Lesson 2.
5. Fontjoy generates font pairings using a neural network trained to find typefaces with:
✓ Correct — Correct. Controlled contrast is Fontjoy's documented pairing principle from Jack Qiao's architecture notes.
Fontjoy targets controlled contrast for complementary pairings. Review Lesson 2.
6. In the recommended AI font pairing workflow, LLM suggestions should be treated as:
✓ Correct — Correct. LLM suggestions are a starting shortlist — real copy testing and human judgment must follow.
AI pairing suggestions are first-pass filters, not final answers. Review Lesson 2's five-step workflow.
7. Variable fonts were introduced in the OpenType specification version:
✓ Correct — Correct. Variable fonts were introduced in OpenType 1.8 in 2016.
Variable fonts were introduced in OpenType 1.8, 2016. Review Lesson 3.
8. FontDiffuser (AAAI 2024) was most significant for its application to:
✓ Correct — Correct. The extreme glyph count of Chinese typefaces made few-shot generation particularly valuable — up to 20,000+ glyphs per typeface.
FontDiffuser addressed Chinese typeface production, where complete typefaces require 20,000+ glyphs. Review Lesson 3.
9. The primary advantage of vector-native font generation (VecFusion, DeepVecFont) over raster font generation is:
✓ Correct — Correct. Bézier path output preserves the structural and metric data that professional font production requires.
Vector output (Bézier paths) is the key advantage — it preserves metric information needed for production. Review Lesson 3.
10. Early beta testing of Adobe InDesign's AI layout feature (2023) found that AI-generated layouts were typographically flat primarily because:
✓ Correct — Correct. The layouts were technically correct but lacked editorial intent — the AI applied hierarchy categorically, not rhetorically.
The problem was missing editorial intent — knowing what should feel urgent vs. quiet. Review Lesson 4.
11. Canva's Magic Design defaults to conservative typographic choices. This is intentional because the tool is calibrated for:
✓ Correct — Correct. Canva's own blog post (2024) confirmed this is a deliberate product decision for their primary non-designer user base.
Canva explicitly designed for non-designers needing fast, functional results. Review Lesson 4.
12. A "widow" in typographic terms is:
✓ Correct — Correct. A widow is a short final line stranded at the top of a new column or page — disrupting typographic rhythm.
A widow is a short final line isolated at the top of a column. An orphan is the first line left at a column's bottom. Review Lesson 4.
13. The WhatFont browser extension identifies web typefaces by:
✓ Correct — Correct. WhatFont reads CSS properties — no image analysis involved. This is why it is more reliable than vision-based tools for web type identification.
WhatFont reads CSS declarations directly — no AI vision involved. Review Lesson 1.
14. When using AI layout tools, prompting specificity should include all of the following EXCEPT:
✓ Correct — Correct. Exact pixel dimensions are an output to be determined, not an input to the creative brief. Emotional register, reading sequence, and visual dominance are the inputs that direct AI layout decisions.
Pixel dimensions are output decisions — let the AI propose them given the intent. Focus your prompts on emotional register, reading sequence, and dominance hierarchy. Review Lesson 4.
15. The fundamental limitation of all AI font pairing and layout tools for professional design work is that they can apply a typographic system but cannot:
✓ Correct — Correct. AI tools are sophisticated automatic typesetters — they apply systems. Authoring the typographic system, with all its intentional meaning, remains irreducibly human work.
The core limitation is authorship — creating the typographic system with its intended meaning. AI applies; designers author. Review Lesson 4.