Navigating the AI Evolution in AEC: From Classic to Contemporary and Beyond

- Posted on

The past decade has been a pivotal transformation within tech across the AEC industry. With a reputation for late tech adoption, AI came into the built world as a major disruptor that revolutionized the way that we build today. As the industry continues to venture further into the fabric of AI, it is clear that its impact is paramount for the industry to continue to navigate the ever-evolving and complex needs stakeholder are demanding. Within this briefing will be a look back to classic AI and its solution offerings, along with a look at what is being used in contemporary AI today.

Want to explore AI and ML in AEC alongside industry leaders?

Register for the Inaugural BuiltWorlds AI/ML Annual Meeting!

The AI/ML Inaugural Annual Meeting is an intimate gathering of thought leaders and innovators from across the AEC industry to discuss trends, challenges, and opportunities arising from the building and infrastructure industry's accelerating data race and turn toward artificial intelligence and machine learning. This special event is limited to participants from 30 companies.

Classic AI

1955

  • Some of the earliest traceable works within this field are set back to a Dartmouth Conference that occurred in 1955 in which the term "artificial intelligence was coined. AI during this period was mainly focused on symbolic methods, such as the Logic Theorist, which was a program that solved mathematic theroums with heurist techniques. Although it may seem simple today, this was a huge milestone within AI, as it showed that computers have the ability to emulate human reasoning and problem solving skills.

1980

  • The 1980s brought a rapid maturation of AI, with rudimentary softwares and methods such as Expert systems and Fuzzy logic making their way into the field.
    • Expert systems harness the power to mimic decision-making abilities of a human with strong expertise within their field. This system holds a knowledge bank that allows it to advise users on how to solve complex issues within their niche.
    • Fuzzy logic is a method of computing that is valued within degrees of truth, rather than just being identified as true or false. This helps to demystify ambiguous input information and provide structured solutions.

1990

  • Following the conclusion of "AI winter", the early 90s was a large surge of funding and research was placed into the field. There was significant progress made within machine learning's algorithms and methodologies, such as the advancements made within neural networks, decision trees, and support vector machines.
  • AI solutions became more widely adopted within the AEC space in the mid-90s. This began its use within the space for risk assessment, project optimization, and cost estimations that have matured greatly within the past decade.

2000

  • What many people commonly think of as classic artificial intelligence began to manifest within cloud-based applications like Microsoft Word and Excel in the early 2000s, followed by its integration into smartphones and tablets in the 2010s.

Contemporary AI

  • The primary difference between classic and contemporary AI is the enhanced ability to incorporate more advanced techniques. Early stages of what is considerd contemporary AI started gaining traction within AEC in the late 2010s and have been rapidly evolving ever since.
  • Advanced AI within AEC holds a diverse set of software tools and tech that incorporates key techniques such as Deep Learning, Natural Language Processing (NLP), and Image Recognition.
    • Deep Learning is a complex branch of machine learning that trains artificial neural networks to mimic human cognition processes. This technique's versatility has lended to its use within a wide breadth of sectors.
      • Use Case Examples: Image/speech recognition, Natural Language Processing, and predictive analytics.
    • Natural Language Processing (NLP) directly focuses on the interactions between computers and human language. This empowers computers with the ability to understand, interpret, and generate human language and natural language data in a way that is both accessible and effective.
      • Use Case Examples: Document analysis and insights, intelligent document management and chat bots/virtual assistants.
    • Image Recognition enables computer systems to identify and categorize specific objects or patterns within a digital image or video. With this process, visual data is interpreted and analyzed through machine learning algorithims and deep neural networks.
      • Use Case Examples: Identifying construction materials, identifying potential risks and current problems, and site monitoring/site safety.

Want to gain a deeper understanding of AEC tech solutions?

Read BuiltWorlds' Tech Specialty Reports

Looking Ahead

As we continue to move further into the tech era, the AEC space is set to be shaped by AI advancements. The future of the industry is project to witness transformational results through the integration of new-age tech. These solutions will continue to strengthen AEC projects, securing enhanced efficiency in terms of cost, time, and sustainability.