Gains:
- Ability to discover form language, ratio and visual balance decisions with artificial intelligence support and compare alternatives
- Ability to incorporate anthropometry and ergonomics principles into the work by having artificial intelligence produce a draft/checklist
- Ability to verify ergonomic and dimensional recommendations of artificial intelligence with real anthropometric data, mockups and user testing
It's no coincidence that a product feels "well-designed." The proportions of the form, the rhythm of the lines, the balance of the masses and how the product fits in your hand or eye are the sum of thousands of small decisions of the designer. In this unit we cover two intertwined topics: aesthetics/form language (how the product looks and feels) and ergonomics (how the product relates to the human body). Artificial intelligence works in very different roles in these two areas. AI in aesthetics is a powerful discovery and comparison tool; In ergonomics, it is just a generator of blueprints and checklists, because real ergonomic decisions are made with human data and physical testing. Confusing the two is the most important caveat of this unit.
Form language and proportion: Aesthetic exploration with AI
Form language is the consistent form decisions that create the visual personality of a product: whether corners are sharp or soft, whether surfaces are flat or curved, whether lines are horizontal or vertical. Proportion is the size relationship of parts relative to each other; A good ratio makes a product look "established", a bad ratio makes it look "makeshift". Designers use principles such as the golden ratio, rule of thirds, and visual weight balance to evaluate ratio.
AI helps here in two ways. First, it allows you to quickly produce and compare different ratios of the same product side by side (e.g. "same lamp, body-to-arm ratio 1:2, 1:3 and 1:4"). Second, it helps you extend a form language consistently across many parts. But beware: AI has no “aesthetic judgment”; it simply reflects popular patterns in the data on which it was trained. You decide which ratio is right for this product, this brand, and this user.
Tip: When making a rate decision, don't ask the AI to "pick the best one"; Ask for "produce the same form with 4 different ratios" and make the choice yourself. AI produces comparison material, you make the judgement.
Ergonomics: here AI is just an aid
Ergonomics is the science of fitting the product to the human body, perception and movement. The diameter of a handle, the position of a button, the angle of a screen, the height of a seat; These are not random, they are based on anthropometry (the science of human body measurements) data. A critical concept is percentile: it shows what percentage of a population is above/below a certain measure. Good ergonomics is generally designed to cover the 5th percentile (small) to 95th percentile (large) user range; designing for a single "average" user leaves out the majority of users.
This is where the edge of AI begins. Ask the AI "what is the average hand width in mm?" When you ask, it gives you a single number; This number may be hallucinated, it is unclear which population it belongs to, and it ignores the percentile distribution. Ergonomic values are taken from reliable anthropometric data tables, not from AI, and are verified in user testing with physical mockup. You can use AI in ergonomics to: make a checklist of what ergonomic factors you need to consider for a product, have you describe a usage scenario step by step and flag risk spots, or get help interpreting an anthropometric data set. But the source of the number can never be AI.
The following table summarizes how the role of AI in aesthetics and ergonomics differs:
Subject
The role of AI
verification source
Form language discovery
Powerful: produces variation and comparison
Designer judgment, brand guide
rate trial
Powerful: side by side alternative
Designer judgment, visual evaluation
Ergonomic size (diameter, height)
Weak: just draft/reminder
Anthropometric data + mockup + testing
Use case analysis
Medium: draft risk list
Real user observation
Accessibility requirements
Medium: checklist outline
Standards + expert + testing
three mini cases
Case 1 — Quick decision with odds comparison. A designer is unsure of the body height/width ratio for a Bluetooth speaker. It asks the AI to produce the same form in ratios of 1:1, 3:4 and 9:16, puts the three side by side and settles on the ratio 3:4 with the customer in 5 minutes. AI was used correctly here: it produced visual material that accelerated the decision, humans made the choice.
Case 2 — Improper ergonomic value. A team designs a pill box for elderly users and asks the AI about the force of the opening tab; AI says "5 Newtons is fine". The team embraces this. However, for a significant portion of elderly and arthritic users, 5 N is too much; the actual target should be much lower, including users with low grip strength. The box cannot be opened during physical prototype testing. Lesson: ergonomic force/measurement values come from actual data of the target population and are verified by testing; AI's odd number is misleading.
Case 3 — Coverage with checklist. When designing a kitchen knife handle, a designer tells the AI to “make a checklist of all the ergonomic and safety factors I need to consider for this product.” AI; It produces a list of 14 items such as grip diameter, anti-slip surface, grip in wet hands, finger protection, right/left hand, cleanability. The designer uses this list as a starting point, extracting the true value of each item from anthropometric data and testing. Lesson: AI expands the scope (reminds you of the factor you forgot), but does not give the values.
Copiable prompt templates
RATIO COMPARISON TEMPLATE (visual AI) “Keep the same form of this product but change only the [torso/arm] ratio: produce three variations — ratio 1:2, 1:3, and 1:4. Keep everything else (material, angle, lighting) the same so I can compare the effect of the ratio.”
ERGONOMICS CHECKLIST TEMPLATE (text AI)"Produce an ergonomics and usability checklist for [product]. Include: grip/grip, reach, force requirement, vision/readability, right-left hand, different sizes and age groups, accessibility. Add a 'what anthropometric data or test should be verified' column for each item. DO NOT GIVE EXACT numbers; what I need to measure/verify tell me."
USE SCENARIO ANALYSIS TEMPLATE (text AI)"The following user uses the following product in the following context: [description]. Write the usage step by step (extends hand, grasps, turns...) and mark possible points of difficulty, fatigue or risk at each step. These are HYPOTHESES to be confirmed by observation; do not make definitive judgments."
ANTHROPOMETRIC DATA INTERPRETATION TEMPLATE (text AI) "I have the following table of anthropometric measurements: [paste data,source]. How do I determine the [handle diameter] range to cover the 5th and 95th percentile user, explain the steps. USE the values in the table; don't make your own number. Specify which echometer I need."
Weak prompt / Strong prompt
WEAK PROMPT: "What should be the ideal size of this handle?"
STRONG PROMPT: "I am designing a hand drill grip. Target user: professional masters, 5th–95th percentile adult hand. Give me an ergonomics checklist: grip diameter, length, surface texture, trigger reach, vibration, wet/gloved grip. Write down what anthropometric data I need to verify for each item and from what mockup test. DO NOT give an exact mm value; I will use the table and test with the mockup."
The weak prompt pushes the AI into a contrived odd number; The powerful prompt directs it to the right job (scope and validation list) and leaves the source of the number to your data+test.
Common mistakes
- “Asking” the AI for ergonomic measurements and processing them directly into the model. These come from anthropometric data and testing.
- Designing for a single "average user". Design that does not cover the percentile range (e.g. 5–95) excludes most users.
- Delegating aesthetic judgment to AI. AI reflects the popular pattern; You choose the right rate for this product/brand.
- Skipping mockup and user testing. A grip that looks good on the screen may be bad in the hand.
- Considering accessibility as an afterthought. Checklist the diversity of sizes, ages and abilities from the beginning.
Caution: Ergonomics is about safety and health; An incorrect force or size value will result in injury, strain, or an unusable product. AI is never a source of measurement in ergonomics; It is a reminder of what you need to verify the most.
In summary
Form/aesthetics and ergonomics require different verification regimes. AI in aesthetics is powerful: it lets you quickly generate and compare language variations of proportion and form, but the “best” is yours to judge. In ergonomics, AI is only helpful: it can produce checklists and scenario outlines, but values such as size, diameter, force etc. come from reliable anthropometric data and are verified in user testing with physical mockup. Cover the percentile range, rethink accessibility, and don't take any ergonomic numbers from AI.
Application task
Choose a hand-held product (any that includes a handle). Give AI a checklist with the “Ergonomics checklist” template. Then research at least three of the metric items on the list (e.g. grip diameter) from a reliable anthropometric source and write down the actual percentile ranges that the AI does not provide. Separately, produce three different ratios of the same product with the "Ratio comparison" template and justify which one you chose and why.
checklist
- [ ] I used AI as a comparison tool for aesthetic/ratio decisions, making the judgment myself.
- [ ] I took ergonomic measurements from reliable anthropometric data, not from AI.
- [ ] I thought it would cover the percentile range (e.g. 5th–95th).
- [ ] I included physical mockup and user testing in my plan.
- [ ] I added accessibility and size/age diversity to the checklist from the beginning.
- [ ] I did not add any ergonomic values to the model without verifying them.