By Don Macdonald, LEED AP, CEM



Not often talked about is an obscure behavioral and computer science religion called Decision Support, a basic form of Artificial Intelligence (AI) and part of understanding ENERGY MANAGEMENT BEHAVIORAL DECISION MAKING.

Will artificial intelligence evolve and make conscious efficiency choices (or lack thereof) for us? Is much of the human “choice” from efficiency already being eliminated, controlled by the ‘man” and dictated by Government regulations and policies?

In the Terminator the robot returned to earth to kill Sarah Connor whose son one day was to become our savior against the machines.

Energy Management Devices

Will Energy Efficiency (EE) become our post-apocalyptic future with Artificial Intelligence (AI) controlling our lives unknowingly surveying our ever movement like in “Terminator”.  In my movie version of “Terminator” my NEST thermostat tracks my preference for 68 degrees with a vengeance, telling me to close the windows, turn off the bathroom fan while it also monitors my movements from the roof deck hot tub to the bedroom down the long hallway in my home. I wonder if that round glass dial is actually an evil eye (thinking HAL 9000 in 2001 Space Odyssey). My refrigerator tells the grocery store when we need to order food, gas grill when I need a new tank, my doctor how much vodka mixer I used this past weekend, toothbrush my dentist how often I brush and my TESLA while making me feel handsome and smart probably notifies the cops I was speeding before I even got there. then to top it all off my government wants to unlock my encrypted Apple iPhone to learn what smut I was watching while listening to Ted Cruz bombast about Trumps hair style.


Decision Support (DS) science makes choices based on our behavior, one half of effective energy management performance. As AI advances, will it embed itself into us? People with heart conditions already have wireless internal monitors called ICD’s which watch for heart arrhythmia’s 24/7 and automatically pop a note to the doctor hopefully saying they didn’t need to shock the person and didn’t go into sleep mode too often. What happens when AI is so commonplace we become un-knowingly dependent on the machine, like in “Terminator”?

In addition to the equipment aspect of making energy consumption efficient (the gadgets, meters and processes) we have the behavioral science. Often times we ask why EE isn’t more common. It’s because of a phenomena called “FODA”. Whats that, you ask? It’s the EE implementation disease many CFO’s and Wall Street types have when relying on EE data that is not well vetted. FODA occurs when we are unable to make decisions without full knowledge of the outcome.

In today’s energy efficiency landscape, most data still remains unsubstantiated, unreliable to the level Wall Street investment firms require to properly manage risk. While FODA is a scary place to be organization’s such as the Sustainability Accounting and Standard’s Board (SASB) and Environmental Defense Fund’s ICP project have begun to strip away these doubts to produce investment grade tools and results. Put another way, a building energy engineer that builds his/her performance baseline without including weather or use her unit consumption normalization is making decisions based on inaccurate and incomplete information, or FODA.

But FODA isn’t the entire picture. We also have FOMO. Which is how you feel when you missed the chance to read this article published on LinkedIn. You feel left out, uninformed, behind the times as it were. This is an example of social media’s affect called “fear of missing out”. When one commercial building in a portfolio of several hundred scores a great Energy Star or LEED rating many people in the other building may suffer from FOMO. FOMO disease forces us to present our best selves without regard to accuracy and repeatable performance. Claiming that you improved upon last years’ energy consumption using only a utility bill and then reporting it in the annual sustainability report is an example of FOMO. Annual reports are rife with sustainability claims not based on repeatable base-lined, properly M&V executed performance standards.

As our dependence on AI grows we need to recognize the programmer behind every software package is a person who is human, may have political, financial, and behavioral views which could be programmed into the software logic. Could we have motives such as greed, vanity, or exploitation found in software without standardization and review processes? Perhaps.


Sensitive to the slightest cumulative environmental changes

Sensitive to the slightest cumulative environmental changes

One of the cumulative effects we are already seeing impact energy efficiency is known as the “butterfly effect”. These separate events, decisions, and bias’ accumulate slowly at imperceptible speed into overwhelming processes which then can spin out of control. When we add FODA and FOMO affects things become difficult to understand where the product begins and behavioral performance ends.

These tiny differences in software efficiency bias’ can cascade into massive changes over time distributed across the GRID with huge results. A well-intentioned computer programmer can make the smallest tweak to my office HVAC cooling. The next thing you know, everyone has personal heaters under their deck. Its subtle, but the potential is there. What happens if the programmer is not well-intentioned? It is the worrisome challenge we face as energy efficiency professionals to improve performance while moving toward automating software controls. We are already moving to automated (AI) systems of remote assessment, there called Building Management Control Systems (BMS’s) and Energy Management Information Systems (EMIS’s).


  • Antonio Gomes Martins

    Nice set of reflections. Speculations like this one are important to make us aware of what may go wrong when we stop questioning the virtues of certain moves that tend to become mainstream.
    I have one remark, though. Although I can understand the role of some exaggerations for the sake of effectively making a point, I don’t think that decision support should be considered a basic form of artificial intelligence. Operations research is much older than artificial intelligence, serves a wide number of purposes and has an enormous potential for useful applications in all sorts of human activities. AI uses OR because OR is very handy and indispensable for many AI approaches. OR allows effective decision support but is a science in itself, far from any dogmatic view of the world. In fact, it is quite the opposite. Even to make a point, it is quite misleading to the average reader to call it an “obscure behavioral and computer science religion”.

    • Don Macdonald


      Your knowledge in AI space is well above mine but I appreciate the comments on the reflections to bridge the competency gap’s between the general reader and the experienced AI operator is an important concern for the future.

      In the current energy efficiency space, engineers are just beginning to include behavioral science value in programming to embed efficiency decisions into controls. We consistently see projects executed using products and hardware without the behavioral component, thus foregoing full efficiency performance potential. We have results which demonstrate that the energy efficiency gains made from equipment are about as equal as those from human behavior modification. This includes everything from lighting, boilers operations, to full building and manufacturing plant processes.

      Standardization of these processes is slowly taking place from different viewpoints. For example ISO 50001 EE standard deals with behavioral management control, whereas ASHRAE 90.1 latest revision includes equipment EE requirements. Several new combined standard is now in the public domain. EDF’s Investor Confidence Project (ICP) is one such example but is in its infancy.

      As these project management processes mature, E,M & V protocols standardize, and behavioral decision making embed in the combined EE space I believe EE products (e NEST, etc.) and processes (how humans adjust and monitor these equipment) will merge and the freedom to choose to be wasteful will decrease, the rise of the machines will have begun.

      Don Macdonald, Managing Director Veritatis Advisors


Leave a comment