by Lauren Schmidt, PLA, ASLA
This post is a summary and recap of the presentation Putting AI to Work: Practical Applications of AI in Landscape Architecture that took place at the ASLA Conference on Landscape Architecture in Minneapolis this past October. This post was written with the assistance of ChatGPT. The full PDF of the presentation is also provided at the end of the article.
Presentation by:
- Benjamin George, ASLA, Utah State University
- Phil Fernberg, ASLA, OJB
- Qing Luo, ASLA, Oklahoma State University
- Lauren Schmidt, PLA, ASLA, Parallax Team
- Tony Kostreski, PLA, ASLA, Vectorworks
- Matt Perotto, ASLA, Hargreaves Jones
Artificial intelligence (AI) has become a transformative force in various industries, and landscape architecture is no exception. This presentation explores the practical applications of AI in landscape architecture, featuring insights from experts in the field.
Understanding Artificial Intelligence and Machine Learning
AI, a concept contemplated for centuries, is the development of computer systems that mimic human intelligence. AI can be categorized into Artificial Narrow Intelligence (ANI) and Artificial General Intelligence (AGI). Landscape architecture benefits from ANI, which includes machine learning, knowledge-based systems, natural language processing, optimization, robotics, and computer vision.
The presentation breaks down the core components of AI, emphasizing machine learning (ML) as a subset that enables computers to learn without explicit programming. Deep learning (DL), a subset of ML using artificial neural networks, allows processing more complex patterns. ML is further categorized into supervised, unsupervised, and reinforcement learning, each playing a unique role in landscape architecture applications.
Text Generators
Generative AI, a powerful facet of AI, involves systems that generate new content, spanning text, images, audio, and virtual environments. Large Language Models (LLMs) like GPT (Generative Pre-trained Transformer) exemplify this capability. The presentation introduces ChatGPT, a chatbot developed by OpenAI, explaining its architecture and applications in landscape architecture.
Applications of ChatGPT in Landscape Architecture
Lauren Schmidt and Qing Luo detail the ways landscape architects can leverage ChatGPT. From answering simple questions about design and cost estimates to undertaking complex tasks such as code summaries and framework development, ChatGPT proves versatile. The presentation emphasizes the tool’s usefulness in brainstorming ideas, looking up design dimensions, obtaining cost estimates, and generating plant lists for specific regions.
Limitations and Best Practices
Despite its capabilities, ChatGPT has limitations, including potential inaccuracies and inability to access the internet. The presentation stresses the importance of verifying information from reliable sources and exercising common sense. A cheat sheet for designers is provided in the presentation, offering guidance on crafting effective prompts.
Image Generators
The presentation introduces various image generators, including Midjourney, DALL-E, Stable Diffusion, and Photoshop. Each platform has unique features, from stylization to accessibility. Benjamin George and Tony Kostreski discuss platforms and access, prompt recipes, section view demos, plan view demos, stylization, constraints, and more.
Getting Started with Midjourney Recipes
The presentation offers practical examples of using Midjourney for idea generation, plant combination studies, seamless patterns or textures, and HDRI backgrounds. Prompt recipes and examples guide landscape architects in effectively utilizing Midjourney for their specific needs.
Limitations
Image generators in landscape architecture have limitations, including a lack of spatial awareness, data constraints in site specificity, variable accuracy leading to potential hallucinations, challenges in user and model comprehension due to linguistic and semantic constraints, monotony in generated outputs, and a current scarcity of “few-shot” capabilities for landscape architecture, though this landscape may evolve quickly.
Conclusion & Resources
Check out the QR code at the end of the presentation for a more comprehensive list of AI tools by design task, put together by Matt Perotto.
Join the Conversation
- What AI tools have you found most practical for your landscape architecture workflow?
- How can we ensure responsible AI development and implementation in our field?
- If you are interested in joining the AI Working Group (part of the ASLA Digital Technology Professional Practice Network), please reach out to propractice@asla.org.
Lauren Schmidt, PLA, ASLA, is a licensed landscape architect and design technology specialist at Parallax Team, a BIM and Practice Technology consultancy that specializes in assisting AEC firms. With 10 years of experience using Revit specifically for landscape design, she leads the landscape effort at Parallax, which includes development of landscape-oriented training and implementation, creation and maintenance of custom project templates, landscape content libraries, and landscape add-in for Revit, FOREground. In addition, she co-chairs the ASLA Digital Technology Professional Practice Network (PPN)‘s BIM Working Group and has a blog, landarchBIM.com, that serves as a resource for landscape architects using Revit and/or Dynamo.