Use of AI in the design process
By Alex Pull; Deputy Head of Year 7, Teacher of Design & Technology, The Grammar School at Leeds
By Alex Pull; Deputy Head of Year 7, Teacher of Design & Technology, The Grammar School at Leeds
As a design and technology teacher, I have always been passionate about exploring new approaches in my practice. In recent years, AI has gained traction in schools, and I am fortunate that my current school embraces its use in teaching. Staff are encouraged to explore various AI tools, and many of us have been using ChatGPT and Copilot for tasks like developing revision materials and providing assessment feedback. This year, I focused on how generative AI could enhance the student design process.
While my experience with generative AI was limited, I recognised its potential. However, I was unsure how to integrate it into design lessons without diminishing creativity or replacing traditional idea generation. A common challenge for students is visualising their designs in 3D. Many struggle with abstract thinking or lack confidence in their drawing. This year, I found that using NewArc.ai helped bridge this gap.
Image 1 – generated using DALL-E inside ChatGPT. Prompt was “A collection of line drawings showcasing modern small house designs, each from a single front-facing viewpoint. The houses display unique modern architectural style.”
Image 2 – generated using DALL-E. Prompt was “A series of line drawings focused solely on the front elevation of modern small house designs, inspired by organic forms. Each house features flowing forms.”
Image 3 – DALL-E generated image showing a brutalist style building.
In Year 9, we run a successful architectural modelling project. Originally developed by my Head of Department during COVID to encourage home-based modelling, we have refined it to align more closely with GCSE coursework and the role of an architect.
The project begins with research on a range of key architects. We provide a diverse core group and encourage students to explore further using websites like Dezeen and our own library resources. While we try to introduce broad architectural styles, students often struggle to differentiate between authentic styles and incorrectly tagged images. To try and address this, I used ChatGPT to generate overviews of ten architectural styles, along with AI-generated images of houses in each style. Though not based upon real buildings, these visuals gave students a clear reference point for their own research.
After research, students sketch simple line-drawn ideas for a small cabin house. We emphasise single-elevation drawings to encourage quick iteration. From here, they transition into modelling, first using SketchUp and then constructing physical models in grey card. This stage helps them translate 2D concepts into 3D forms. Scale is gradually introduced using a 1:100 ratio, with figures for reference.
Image 4 – Student design work before any AI use. The buildings are drawn from a single viewpoint and completed quickly.
Once students refine their models, they transfer their designs into 2D Design. Here, they focus on accurate scaling, ensuring correct proportions for doors, floors, and overall dimensions. Once complete, they laser-cut the parts and reassemble the models using PVA glue. This iterative process helps them identify and correct errors before cutting the final model in MDF. Students then place their models on standard-sized plots, considering terrain and external features like driveways or greenery.
Image 7 – Final MDF model of A frame house.
Image 6 – Student drawn image showing a circular house design. Lots more variation in the designs, with some clear errors.
Image 5 – Student drawn image of and A Frame house and the results from NewArc.ai. There is some variation in the images, but shows the building in a final form.
NewArc.ai has been a valuable tool in this process. Students crop 2D elevations from their sketches and final drawings, and I use NewArc.ai to generate renders based on their designs and additional prompts. Each drawing produces four images, with the best ones shared with students. The results often reveal different material applications or alternative configurations, enhancing students’ understanding of their concepts. These AI-generated images are incorporated into their final project presentations, alongside their own design work.
Integrating AI into design lessons has been really interesting, and student feedback is very positive. Tools like NewArc.ai provide students with valuable insights into their own design capabilities, helping them overcome creative blocks and visualise their designs more effectively. This experience has reinforced my belief that AI can support, rather than replace, creativity in education.