Artificial intelligence has been in the headlines for the past several months. People have expressed concerns that it will take their jobs, alter the operation of the technology and investment industries, and even make people lazier by reducing the need for education and training to solve complex issues. How much of this worry is justified, and how much is overhyped catastrophizing? Let’s explore what AI is and what it means for the roles our NEBB NEBB-certified professionals, Technicians, and Firms play in the building industry:
What is AI and What Can It Do?
Before diving too deeply into the impact that artificial intelligence is having on the industry at large, we need to define what exactly AI is and how it operates at present. The form of artificial intelligence that is publicly available today differs significantly from what you may have seen in futuristic sci-fi movies. Current AI technology is not yet at the stage where it can be implanted into robots that are then released to better society by independently making decisions. Rather, current technology has developed to the point where it encompasses several different tiers. Let’s attempt to break down each of the levels of AI, simplified as much as possible, as it relates to the building industry:
Level 1:
You may have heard of the first level of AI, machine learning, which occurs when a computer uses an algorithm to analyze categorized and labeled data, allowing the computer to understand what it is analyzing. The computer then uses that dataset to make decisions. The limitation at this level is that a person must categorize the data, indicating what information is relevant to the decision or calculation the computer is attempting to make. However, once someone sorts through the data and inputs it into the machine, it is the computer itself that processes the information and provides an output.
Level 2:
The next layer of AI is called deep learning. At this stage, the computer can pull data from various sources and categorize the information independently. Where it previously required human input to identify which data belonged in which decision box at Level 1, at the deep learning level, the computer can analyze a large pool of unsorted data and determine for itself which pieces of information to use for a specific decision. This reduces, or even eliminates, the need for human interaction in sorting and categorizing the information. This is the core of how current AI models, like ChatGPT, operate. They use the entire internet as their data pool and can sift through that information, automatically categorizing what is relevant to the question being asked before providing a response. However, it is the manner in which these models respond that determines whether they reach the third—and current— level of artificial intelligence.
Level 3:
Once machines could sort through pools of data and make decisions, it became necessary to find a way for them to communicate these decisions in an understandable way. Large language models—essentially vast collections of written works demonstrating various uses of language—were given to computers to help them learn sentence structure and composition. Over time, this led to AI models being able to compose original sentences and communicate with people in written language. Similar processes were applied to pictures, videos, and artwork, allowing AI to create original content in these mediums. This is the level at which AI currently operates, known as generative AI. At this level, all the abilities of Levels 1 and 2 combine, enabling the computer to produce outputs comprehensible to humans. This also means that humans can input queries in common English, and the computer can interpret and provide a logical response.
As an example, you may have heard stories of computer programmers losing their jobs en masse as AI reduces the need for people in that industry. This concern is valid; AI has become adept at translating written prompts into computer code across various programming languages. For this reason, if you have children in school taking programming classes, you have likely heard about the rules their instructors have implemented to guard against the use of AI to complete homework assignments. Students can input the exact word problem from a homework assignment and have an AI model generate an entire executable program in a matter of seconds and at no cost!
Does this mean we should immediately hit the panic button and flee before the machines take over? Not quite. Here’s a counterbalancing point: artificial intelligence can scour the internet to quickly review thousands of articles on a given topic, consolidating what would otherwise take hundreds of hours of research into a few paragraphs. However, using it as an advanced research tool has revealed limitations. When gathering and summarizing information, AI has not yet progressed to the point where it can reliably distinguish between credible sources and misinformation. Therefore, while AI can use programmed models to simulate deductive reasoning and make accurate decisions based on reliable information, it fails if the provided information is inaccurate but appears relevant. In simpler terms, AI is subject to the adage, “garbage in, garbage out.” It relies on good data to function effectively, and the internet is hardly a repository of solely reliable information.
You may also wonder whether AI has the ability to learn from past engineering documents to create and produce new systems that fit within a given footprint. Surprisingly, significant progress is reportedly being made on this front. Software developers are diligently working to build libraries of engineered drawings that AI can use as source material to generate new designs for installation in buildings. This capability of AI, like all developing technologies, is improving over time but is still not foolproof. AI models excel in specific tasks, such as dimensioning and checking load and sizing calculations, but skilled engineers are still needed to provide input and ensure that the generated plans will work effectively.
By extension, there remains a demand for skilled tradespeople who understand how building systems function and interact with one another. These professionals need to possess a solid grasp of basic design principles, as well as an understanding of how various devices operate and interact when installed in a given space. AI is advancing rapidly in the building automation space, but at this point it holds a complementary role rather than one of replacing a need for human capital altogether. While it is imperative that we take note of where AI is now, and where can we expect this technology to go in the near future, it is just as important to understand that it is the highly regarded work of NEBB Professionals, Technicians and Firms adhering to NEBB Procedural Standards, which set us apart and help us showcase the meaning of integrity itself.