What happens when we replace Natural Light with Artificial Light? 

When I talk about the environment influencing our health, I mean something more fundamental than pollution, access to green spaces, or whether we live somewhere considered healthy. I'm talking about the physical conditions surrounding us and the information those conditions provide to our biology. Light, darkness, temperature, gravity, and the other forces we're continuously exposed to influence how our bodies function. We've developed within those conditions over an extraordinarily long period of time, and although we've become remarkably good at manipulating our surroundings, that doesn't necessarily mean we understand the biological consequences of changing them.

One of the clearest examples of this is something most of us probably give very little thought to: the light we're exposed to every day. Over the past several decades, governments and manufacturers have substantially changed the kinds of light used to illuminate our homes, workplaces, streets, and public spaces. What initially appears to be a relatively straightforward decision to improve energy efficiency becomes considerably more complicated when we begin examining what light actually is, how our bodies respond to it, and what was considered when governments decided to change the way we illuminate our surroundings.

To understand the problem, we first have to understand that light does considerably more than allow us to see. Light is electromagnetic radiation, and different wavelengths interact with biological systems in different ways. We've known for a long time that sunlight influences vitamin D production, but that's only one of the better-known examples. Specialized photoreceptors in the retina detect light and communicate with regions of the brain responsible for regulating our circadian rhythms. These rhythms help coordinate sleep and wakefulness, hormone secretion, body temperature, metabolism, and the activity of tissues throughout the body. Researchers are also investigating how particular wavelengths of red and near-infrared light interact with cellular processes, including those involved in mitochondrial function.

This is why I think there's a reasonable argument for considering light a biological nutrient, even if it doesn't fit the conventional definition of something we consume through food. We require exposure to particular forms of light, at particular times and intensities, for normal physiological regulation. The absence of appropriate light can have consequences, as can exposure to light at the wrong time. And like other environmental inputs, the effects depend on the characteristics of the exposure and the biological state of the organism receiving it.

Yet when we purchase a light bulb, we're generally told to evaluate it according to watts, lumens, and color temperature. Watts tell us how much electrical power the bulb consumes. Lumens describe the quantity of visible light, weighted according to the sensitivity of human vision. Color temperature describes whether the light appears warm or cool. These measurements are useful for their intended purposes, but none of them adequately describes the complete physical composition of the light or its biological effects.

Sunlight, firelight, and incandescent bulbs share a physical process called incandescence, in which heated matter emits radiation. An incandescent bulb produces light by heating a tungsten filament until it glows. Its emission follows a relatively continuous spectrum, with considerable energy extending into the infrared region. A conventional white LED works differently. It typically uses a semiconductor to produce blue light, some of which is converted into longer wavelengths by a phosphor coating. The result can appear visually similar to incandescent light while having a substantially different spectral distribution.

This happens partly because of how our visual system works. Our eyes and brains can produce the perception of the same color from different combinations of wavelengths. Two light sources can therefore look nearly identical even though the actual radiation reaching our bodies is different. These differences become much clearer when we look at their spectral power distributions. Incandescent light produces a relatively smooth curve, while conventional fluorescent lamps and many white LEDs show pronounced peaks. They can deliver comparable visible brightness while exposing the body to different wavelengths at different intensities.

I find this particularly interesting because the assumption behind much of our lighting technology seems to be that if the light appears the same, and provides an equivalent amount of illumination, it can reasonably be treated as the same thing. But we've already established that light has biological effects extending beyond vision. If those effects depend on wavelength, intensity, timing, and duration, then why would we assume that two sources producing the same perceived brightness are necessarily interchangeable from a biological perspective?

The concern also extends to what's missing from the light. Much of the public discussion about LEDs focuses on their blue-light emissions, particularly in relation to sleep and circadian regulation. But there's also the relative absence of infrared radiation from conventional LEDs and the emerging research into biological responses to red and near-infrared wavelengths. That research doesn't establish that ordinary incandescent lighting provides a therapeutic dose of infrared radiation, or that the absence of infrared from an LED is necessarily harmful. However, it raises another question about the criteria we're using to evaluate these technologies. If we judge lighting almost entirely by the visible illumination produced per unit of electricity, we're already excluding much of the electromagnetic spectrum from consideration.

We can see another part of this problem by looking at the natural cycle of light and darkness. For most of human history, daylight and darkness were environmental conditions we couldn't simply switch on or off. The intensity and spectral composition of sunlight changed throughout the day, and darkness created a very different environment at night. Our circadian systems developed within those recurring patterns and continue to use environmental signals to organize physiological activity.

Today, we can spend the majority of our waking hours inside buildings with relatively little natural daylight, then continue exposing ourselves to artificial illumination well after sunset. We can maintain relatively constant light levels throughout the day regardless of the season, and we can choose lighting that appears warm or resembles daylight without necessarily reproducing the spectral characteristics of either. The body still responds to the physical information it's receiving, even when that information has become disconnected from the natural conditions that historically organized our physiology.

This is already relevant to the relationship between light exposure, circadian timing, sleep, and metabolic regulation. Light reaching the retina helps synchronize the central circadian clock, which in turn participates in coordinating rhythms throughout the body. If those signals are repeatedly mistimed, the consequences may extend into other physiological systems. So when we talk about sleep problems, metabolic dysfunction, fatigue, or other health issues, I think it's reasonable to ask how much of the problem might be influenced by the environment in which those systems are operating.

The implications go beyond human physiology. Research on artificial lighting in the natural world has examined changes in animal migration, foraging behavior, predator-prey relationships, reproduction, and other biological activities. Different organisms respond to light in different ways, and introducing artificial illumination into an environment can alter those responses. Some species benefit from the changes while others are disrupted by them. The broader point is that light helps organize biological activity throughout ecosystems, and changing that light can have consequences that aren't obvious simply by looking at how well an area is illuminated.

With all of this in mind, the decision to phase out incandescent lighting becomes much more interesting.

In December 2008, the European Union moved toward phasing out conventional incandescent bulbs, with regulations introduced in 2009 and restrictions taking effect over the following years. The decision was justified primarily through energy efficiency and projected reductions in electricity consumption and carbon emissions. Estimates suggested the transition could save approximately 40 terawatt-hours of electricity annually and prevent around 15 million tonnes of CO₂ emissions. The United States followed a related trajectory through the Energy Independence and Security Act of 2007, which established new federal lighting-efficiency requirements. Subsequent regulations introduced a minimum standard of 45 lumens per watt for general service lamps, with enforcement fully phased in by August 2023, effectively removing most conventional general-purpose incandescent bulbs from ordinary retail sale. In April 2024, the Department of Energy finalized additional efficiency requirements for newly manufactured or imported general service lamps, with compliance scheduled for 2028.

Now, there's a legitimate physical basis for saying that LEDs use substantially less electricity than incandescent bulbs to produce comparable visible illumination. But that observation alone doesn't establish the total economic, environmental, or biological benefit of replacing one technology with another. The savings frequently advertised to consumers are qualified as up to a particular percentage, and the actual savings depend on the bulbs being compared, how they're used, and what happens to overall electricity consumption. A reduction in the energy required to operate an individual bulb doesn't automatically translate into the same reduction in a household's energy bill, particularly when electricity prices, other household demands, and lighting behavior are changing at the same time.

There's another complication worth considering here, because making artificial lighting cheaper to operate may also encourage us to use more of it. We can install lights in places that previously remained dark, illuminate larger areas, increase brightness, or leave lights operating longer because the electricity required to do so has become less expensive. Researchers refer to this as the rebound effect, where improvements in efficiency can be partially or potentially completely offset by increased consumption. In a 2017 study published in Science Advances, researchers examining satellite observations found that the Earth's artificially illuminated outdoor area expanded by approximately 2.2% annually between 2012 and 2016, while areas that were already illuminated also became brighter. A 2023 study published in Science, using more than 51,000 citizen observations of the night sky between 2011 and 2022, estimated that visible-band sky brightness increased by approximately 7–10% per year. Those findings don't establish that the additional lighting has eliminated all the electricity savings achieved through more efficient technology, but they do show that improved lighting efficiency hasn't prevented the continued expansion of artificial illumination.

And that introduces another consequence that doesn't necessarily show up in the calculations used to justify the transition. Even if the electricity required to illuminate a particular area decreases, the biological consequences of extending illumination into places and times that would otherwise be dark can continue increasing. This is particularly apparent in studies of nocturnally migrating birds, which use environmental information to navigate and orient themselves. In a 2017 study published in the Proceedings of the National Academy of Sciences, researchers examining the Tribute in Light installation in New York City found that migrating birds became concentrated around the powerful light beams, reduced their flight speeds, and began following circular flight paths. When the lights were extinguished, the behavioral disruptions disappeared. The experiment involved unusually intense lighting, but it provides a direct example of how artificial illumination can alter animal behavior, and how removing that exposure can restore more normal conditions. A separate 2023 study in Nature Communications found that artificial skyglow was a strong predictor of where migrating birds stopped during their journeys across the United States, raising additional concerns about the consequences of expanding nighttime illumination.

This raises a problem with evaluating environmental progress primarily through energy consumption. We may reduce the electricity required to produce a given amount of light while simultaneously increasing the amount of artificial light present in the environment. From the perspective of electrical engineering, the technology is more efficient. From the perspective of an organism that depends on darkness, the environment may have become considerably more disruptive. Both outcomes can occur at the same time, and a policy focused primarily on electricity consumption may do very little to account for the second.

It also leaves the question of total environmental impact. LEDs require materials, electronic components, manufacturing processes, transportation, and eventual disposal. Incandescent bulbs also have environmental costs, including the electricity required during their operation. A meaningful comparison needs to account for the complete life cycle of each technology rather than concentrating exclusively on one measurement of efficiency. We would also need to establish whether the projected reductions in electricity consumption and emissions were actually achieved, and whether the assumptions used to justify the policy continued to hold after the transition.

What I find particularly concerning is that the European legislation itself included criteria extending beyond energy consumption. Among other things, it required consideration of product functionality, human health, environmental safety, affordability, and life-cycle costs. People with light-sensitive medical conditions challenged the transition because the replacement technologies caused symptoms that they hadn't experienced under incandescent lighting. Organizations representing people with migraines, lupus, and other conditions argued that removing access to incandescent bulbs could make homes and public spaces more difficult for some people to occupy.

Some of these experiences remain difficult to explain mechanistically, and the evidence supporting different proposed causes varies. But the larger regulatory question remains. How was it determined that the available replacements were sufficiently equivalent when the physical characteristics of the light were different and there were still important gaps in our understanding of how those characteristics affected human physiology?

Another problem emerged in the way some of these questions were handled. Scientific committees assessing the European legislation were primarily reviewing existing research, rather than commissioning the additional research necessary to investigate unresolved questions. This created a situation where insufficient evidence of harm could be interpreted as reassurance, even when the evidence was insufficient partly because the appropriate studies hadn't been conducted. I'm not suggesting that every hypothetical risk needs to be disproven before a new technology can be adopted. But when governments introduce legislation that effectively removes a widely used product and directs entire populations toward alternatives, I think the burden of demonstrating the full benefits and consequences of that decision should be considerably higher.

There is also an economic and political dimension to the story. The incandescent phase-out expanded through international initiatives involving governments, environmental organizations, and major lighting manufacturers. The United Nations Environment Programme's en.lighten initiative, for example, was established as a public-private partnership involving companies such as Philips Lighting and OSRAM, with the goal of accelerating the transition away from incandescent bulbs.

Those relationships don't necessarily establish wrongdoing, but they should invite scrutiny of the assumptions behind the policy and the interests involved in promoting it. Manufacturers stood to participate in a massive global replacement of lighting technology, while governments and environmental institutions could point to the transition as a measurable achievement in their efforts to reduce energy consumption and emissions. Meanwhile, many of the possible biological consequences were substantially harder to quantify, and the people most sensitive to the changes were left arguing for access to a technology that had previously been commonplace.

By the time this transition became part of everyday life, incandescent bulbs were routinely described as wasteful, inefficient, and environmentally irresponsible. Those descriptions weren't entirely invented, but they emphasized particular characteristics of the technology while leaving others largely outside the discussion. The argument became increasingly difficult to challenge without appearing opposed to environmental protection itself. A former adviser to the European Environment Commissioner even questioned whether the environmental benefits of some policies were being confused with their political advantages.

I think this brings us to a larger problem with how we define progress, particularly when it comes to the environment and human health. We tend to evaluate new technologies according to the particular problem they were designed to solve. If the goal is to reduce electrical consumption, then a technology that produces more lumens per watt appears to be an improvement. If the goal is to make food cheaper, more convenient, and available throughout the year, then industrial food production can appear to be an improvement. If the goal is to reduce the physical effort required to travel or perform daily tasks, then transportation and automation can appear to be improvements.

And in many respects, they are improvements. But those technologies also change the conditions we're exposed to, the behaviors required of us, and the physiological demands placed on our bodies. Those effects don't necessarily appear in the measurements used to determine whether the technology was successful.

We've created a world in which we can control an extraordinary number of environmental conditions, often according to economic, technological, or political priorities. We can eliminate much of the movement once required by daily life, maintain constant temperatures, extend daylight indefinitely, and eat foods that would historically have been unavailable in a particular place or season. Each of those developments can be defended according to some measurable benefit. But their effects accumulate within the same organism, and the organism doesn't experience them as separate policy decisions or technological achievements.

This is also why I think we need to look further upstream when trying to understand health. We spend an enormous amount of time studying what happens inside the body after something begins to go wrong. We examine glucose regulation, hormones, inflammation, sleep disorders, metabolic dysfunction, and the pathways associated with disease. Those are legitimate places to investigate, but they don't necessarily explain the conditions that produced the biological state we're examining.

If light exposure helps regulate circadian physiology, then the light environment becomes relevant to understanding the downstream effects of circadian disruption. If movement provides mechanical information that tissues need to maintain their function, then the physical demands of the built environment become relevant to musculoskeletal health. If food availability, composition, and timing influence metabolism and behavior, then the food environment becomes part of understanding those responses as well.

The incandescent bulb is one example of how an apparently beneficial change to our environment can raise questions extending far beyond the original purpose of the technology. It also shows how difficult those questions become when the available evidence, the measurements being used, and the incentives surrounding a decision are concentrated on a relatively narrow set of outcomes.

I'm not arguing that we should abandon electricity, artificial lighting, or every other technological development that has changed how humans live. We've gained considerable freedom from our ability to modify the environment, and there's no reason to assume that every ancestral condition was optimal or that every modern departure is harmful. I am questioning whether we have become too comfortable assuming that technological efficiency and convenience are reliable measures of improvement without adequately considering how those changes influence the biological systems that have to live with them.

If health is the capacity to function, adapt, and live well, then the conditions in which that capacity develops and is maintained have to be part of how we evaluate progress. That includes asking what a technology changes about our daily physical experience, how the body responds to those changes, what we actually know about the longer-term effects, and whose interests or priorities determine which outcomes are considered important. In the case of artificial lighting, we have an example of a major environmental change that was evaluated predominantly through energy consumption even though the light itself has biological functions extending far beyond illumination.

And I think that's where the discussion of health needs to continue: examining the physical conditions we've created, the reasons we've created them, and whether those conditions are actually compatible with the physiological requirements of the people living within them.


Further reading: Anna Levin, Incandescent: We Need to Talk About Light (Saraband, 2019).

Previous
Previous

What happens when we leave Earth?