Gains:
- Ability to integrate AI in place and within boundaries at every stage of the workflow, from concept to implementation project
- Ability to establish office-scale prompt library, verification protocol and privacy policy
- Ability to design a control culture that maintains quality, responsibility and customer trust in AI-supported production
Module Exam
1. An architectural office wants to use AI to speed up the final sizing of load-bearing column sections. Which is the most correct approach?
- A) Using AI only for pre-screening and checklisting; Determining the final section with the calculation and approval of a competent civil engineer ✔
- B) Processing the sections given by AI directly into the project
- C) Accepting without calculation if the AI gave the same result more than once
- D) Skipping engineer approval and speeding up the process
Description: Structural system sizing is a safety-critical decision in terms of life safety. AI can be used for pre-screening, checklisting and communication; However, the final section and reinforcement decision must be made with the calculation and approval of a competent civil engineer in accordance with the current regulation. AI output does not replace this approval.
2. What is the safest way to avoid getting a fake (hallucination) value when you ask the AI about the zoning status and construction conditions (TAKS/KAKS, drawing distances) of a land?
- A) Trust the rate given by AI and start the project
- B) Confirming TAKS/KAKS and towing distances from the official zoning status document and current plan notes of the competent administration ✔
- C) Ask the same question again with different words and take the average
- D) Proceeding based on a forum comment
Description: Although the language model knows zoning terms, it can fit rates and draw distances for a particular parcel. This information should only be confirmed from the official zoning status document of the competent administration (municipality) and the current zoning plan/plan notes.
3. What is the best attitude when presenting to the customer an impressive concept render that you produced with a tool that produces images from text?
- A) Presenting the visual as a project that can be implemented exactly
- B) Hiding the scale and structure problem from the customer
- C) Clearly stating that the visual is a concept/atmosphere study and further verifying the applicability ✔
- D) Dimensioning and presenting the visual to replace the technical drawing
Description: AI concept images convey atmosphere and ideas; It does not guarantee scale, structure and material feasibility. It is essential for honesty to clearly state that the visual is a concept study, not to confuse it with the real project geometry, and not to create misleading expectations.
4. What is the most critical check before running an AI-written Dynamo/Python script on your BIM model?
- A) Check that the code does not cause any errors and run it on the entire model.
- B) Relying on the output to look good on the screen
- C) Running the code directly in the main model without reading it
- D) Reading the code and testing it with a small selection on a spare model copy and verifying the result on the model ✔
Description: AI code may assume the wrong parameter name, bulk delete elements, or silently overwrite values. The code should be read, tested with small selection on the spare/copy model first and checked to see if the result is as expected on the model.
5. What is the most appropriate use when you summarize the result of a daylight simulation to AI?
- A) Summarizing the results of the simulation software by giving input to AI, clearly stating the assumption and method ✔
- B) Asking the AI to calculate the daylight multiplier value itself
- C) Using the numbers produced by AI instead of the official calculation
- D) Putting the result in the report without confirming it with accredited software
Explanation: AI does not perform the numerical simulation itself; Summarizes and reports the results produced by accredited/verified software. The correct use is to use AI only as an interpretation and writing tool, clearly stating the input assumptions, boundary conditions, and method.
6. What is the best approach when converting the field program (list of requirements) into a layout variant with AI?
- A) Translating the initial layout given by the AI directly into the application plan
- B) Taking the variants as inspiration and hand-validating them according to square meter, circulation and regulation restrictions ✔
- C) Postponing escape distances to think about later
- D) Not checking accessibility rules because AI knows them
Description: AI is fast at generating layout ideas, relationship diagrams, and variants, but may violate critical rules such as square footage, escape distance, and accessibility. It is necessary to take the variants as inspiration and verify each one according to the net square meter, circulation and regulation constraint.
7. When outputting a bill of quantities with AI, what is the most critical step before trusting the output?
- A) Relying on the table to appear properly formatted
- B) Counting the pose numbers correctly because AI gives them
- C) Cross-verify quantities with independent calculation, actual project measurements and current unit prices ✔
- D) Putting the total amount into the bid as is
Explanation: AI may confuse the unit in the footage (m² with m³), add the area incorrectly or hallucinate the exposure number. It is essential to cross-validate the result with an independent calculation, actual project measurements and an up-to-date unit price list.
8. What is the best way to avoid greenwashing when researching a material's sustainability/embodied carbon claim with AI?
- A) Accepting the AI's repeated marketing claim
- B) Choosing the source that gives the lowest carbon value
- C) Confirming the claim from EPD, independent certification and recognized databases ✔
- D) Relying on a single sentence on the manufacturer's website
Description: AI may be trained in product marketing language and repeat unsubstantiated environmental claims. The actual value must be confirmed from the manufacturer's EPD (Environmental Product Declaration) document, independent certification and recognized databases.
9. What is the best behavior in terms of confidentiality when giving the drawings and data of your client's unpublished competition project to a cloud-based AI tool?
- A) Anonymize context, clear identifiers, and use a policy-compliant tool where data will not be used in education ✔
- B) Uploading all drawings and customer name as is
- C) Using a free tool without reading the privacy policy
- D) Assuming that location information must be given for the quality of the result
Disclosure: Unpublished project, location and client information is trade secret and copyright sensitive. It is necessary to anonymize the context, clear identifiers such as location/name, and choose an institutional and policy-compliant tool that guarantees that the data will not be used in education.
10. What is the best approach if the recognizable style of a contemporary architect you are inspired by is imitated exactly in an image produced by AI?
- A) Copy the style exactly and present it as an original product
- B) Using the visual as internal inspiration and transforming it into an original design language and being transparent about the source ✔
- C) Assuming that the copyright issue does not apply to architectural images
- D) Never telling the customer who owns the style.
Explanation: Copying another architect's registered/recognizable style verbatim and presenting it as an original product is risky in terms of copyright and professional ethics. It is necessary to use the visual only as internal inspiration and transform it into an original design language and keep the source transparent.
11. What is the basic structure that should be established to maintain quality and responsibility in an AI-supported architectural office?
- A) Every employee uses unsupervised AI in their own way
- B) Importing AI outputs directly into the project without saving any
- C) Establishing a common prompt library, verification protocol and privacy/copyright policy and closing security-critical outputs with competent approval ✔
- D) Perform verification only once at the end of the project
Description: Distributed and uncontrolled use of AI magnifies the risk of errors and privacy. Establishing a common office-scale prompt library, verification protocol for each output type, and a clear privacy/copyright policy; It is essential to close each security-critical output with competent approval.
12. When you asked the AI about a fire escape distance rule, it gave an exact number. What should you do before using this number in the project?
- A) Apply directly as AI gives a clear number
- B) Replacing the number with the value you remember from another project
- C) Rounding the escape distance regardless of the source given by the AI
- D) Confirm the value from the official text of the current fire/building regulations and have the design approved by a competent expert ✔
Description: Escape distances are safety-critical and regulation dependent for life safety; AI may remember this value incorrectly or give it from a different country's regulation. The value must be confirmed from the official text of the relevant fire/building regulations in force and the design must be approved by a competent expert.
13. The most efficient way to use AI at the concept stage is to give it what?
- A) Give clear and structured constraints such as land context, climate, program, budget and design intent ✔
- B) Just give a one-sentence request such as 'design a beautiful building'
- C) Producing many images and choosing one without giving any context
- D) Let the AI assume the constraints itself
Explanation: The output quality of AI depends on the input quality. The output works when you make clear constraints such as site context, climate, schedule, budget and design intent; Vague one-sentence requests produce general and superficial results.
14. You have drafted a technical specification with AI. What is the most correct attitude towards this draft?
- A) Putting the draft in the tender dossier as it is
- B) Use the draft as a framework and review and correct each clause and standard reference by a competent expert ✔
- C) Not checking the standard numbers because AI gives them
- D) Accepting material descriptions without looking at the project conditions
Explanation: The technical specification creates legal and technical liability as an annex to the contract. AI can quickly produce a draft skeleton; However, each item, standard reference and material description must be reviewed by a competent expert and corrected according to the applicable standard and project requirements.