CO2 Out Of Sight, Not Out Of Mind

As the CO2 concentration in the atmosphere peaked in 2016, limiting the global temperature increase to only 2 degrees might become unattainable. Although the energy transition is well under way, phasing out fossil fuels might take many decades due to the growing energy demand worldwide [1]. The good news is, there might be a surprising solution which lies where the fossil fuels come from – deep underground. Carbon Capture and Storage (CCS) technology can deliver 12% of the cumulative reduction of emissions required by 2050. However, there are only 16 large-scale projects currently in operation and the governments are still to address multiple gaps in the regulatory frameworks across the world. One of the stumbling blocks is the widespread perception that this new technology comes with the risks that are just too high and hence does not deserve to be developed further. But how objective are we about the risks involved?

For most people, the ill-effects seem too distant for climate change to meaningfully impact their behavior.

The climate on our planet that allowed our civilization to flourish is a classic example of a public good: everyone enjoys it irrespective of whether they spend their money and effort to maintain it. This affects our decision-making, because it gives rise to a free rider problem. That is, people are not necessarily willing (or able) to pay their fair share to incur the costs of maintenance. Moreover, change in the climate can barely be noticed within one generation. Thus, for most people, the ill-effects seem too distant for climate change to meaningfully impact their behavior. This is why it is so important that we all recognize the problem and look for solutions together. In the case of CCS, the negotiations with governments and the public have proven very difficult.

First, let’s look at how CCS works. This technology allows to do exactly what the name says: capture, transport (if required) and securely store CO2, usually several kilometers underground, at a suitable location chosen by geologists. The gas is then pumped into a reservoir with an impenetrable cap rock – often the same reservoir that was keeping the oil and gas in place for millions of years. As of now, this is the only proven technology that can be used on a global scale to reduce emissions. There is no doubt that the transition to renewables and the search for technologies that deliver higher energy efficiency should continue, but there are many production processes where fossil fuels cannot be substituted yet, such as production of steel, iron and cement, to name a few. CCS can be an important “interim” technology capturing these CO2 emissions until we find out how to produce these materials in a cleaner way.

Like any technology, it comes with risks, the main one being a CO2 leakage. An abrupt leakage might be dangerous for the local community, because high concentrations of CO2 pose health risks. Research shows that the probability of such an event is extremely small and it is very unlikely to cause harm to people or local flora and fauna [2]. Nonetheless, the technology is still perceived as unacceptably risky by many, which might stem from the behavioral biases we possess.

Studies have shown that people are prone to omission bias [3]: that is, they favor inaction allowing harm to be done instead of an action that might cause harm. Baron (1986) notes that “omission bias seems to arise from a basic view that what is to be avoided is the direct causation of harm” [4]. He demonstrated this effect in parents’ decision to vaccinate children: many opt out thinking that vaccination itself poses risks for the child and hence they do not want to potentially cause harm by their actions, even though statistically the advantages of vaccination far outweigh the risks. CCS is perceived in a similar way. With respect to this technology, we seem to have two options:

build a CCS and take on the (however small) risk of potential leakage


do not build anything and face the consequences of climate change.

The latter prospect seems far from ideal, but the public might favor it simply because they want to avoid causing any harm at any cost. The fact that environment is a public good aggravates this perception, because the initiatives to reduce CO2 emissions might not benefit directly to the country or community introducing them and will be effective only if implemented on a global scale. In a way, this is similar to the prisoner’s dilemma, because cooperation of all the countries and communities is not guaranteed, so each one of them is inclined not to take the risk of adopting CCS on their own. That is why using a technology that implies that potential risks are localized but the benefits are shared requires us to recognize this bias and coordinate our effort globally.

Another problem is that people prefer to focus on reducing a smaller risk to zero rather than significantly reducing a bigger risk. This is known as zero risk bias. For CCS, this means that communities and politicians would oppose building a CCS facility (i.e. reducing the risk of leakage to zero) rather than contribute to a reduction of a greater risk of global warming. In this case it is important to frame the options clearly and do an objective assessment of all the risks involved. On the one hand, we have a risk of CO2 leakage. The estimated probability of this happening is miniscule [5] and is very unlikely to cause harm to living organisms. On the other hand, if CO2 emissions are not curbed, climate change will certainly worsen.


The AI Governance Challenge

It is also worth mentioning that people usually prefer to reduce man-made hazards rather than similar hazards coming from natural phenomena. In their research on public opinion on cleaning up hazardous waste, Baron et al (1993) [7] found: “A few subjects think that human-caused waste is more important to clean up than otherwise identical natural chemicals”. CCS is a similar case: the deposits of CO2 can be found in many places on the planet, but they never spark discussions about their hazards. Nevertheless, natural CO2 deposits might be much more dangerous than man-made CCS because their locations have not been specifically chosen by geologists, so the cap rock might not be strong enoughto prevent the gas from leaking. This is a perfect example of naturalistic fallacy– the tendency for people to view a phenomenon that occurs naturally as more positive compared to a similar man-made phenomenon – which in this case complements and intensifies the zero risk bias.

Fig 1: Areas of natural CO2 deposits globally

CCS certainly has its risks, so technical studies and tests should still be done and advanced regulatory frameworks should be developed to ensure this technology can be employed safely. For example, we could start by storing CO2 in the remote areas where potential leaks would not affect anyone until we gather enough data. CCS does not claim to be the ultimate solution that would stop climate change once and for all, but it seems to be the necessary stepping stone to a net zero emission world. Meanwhile, scientists are working on some amazing technologies for the longer term — such as turning CO2 into stone. But it will take us decades of research before we are able to use those on the required scale. So, for now, it is important to take stock of the technological ability we do have, and ensure we check our biases in making objective decisions that are best for the planet.

New Year’s Resolutions: Why We Make Them and How to Keep Them

It’s been the first few weeks of January, and someone inevitably asks: Do you have any New Year’s Resolutions? After a moment of thought, you are instantly reminded of your loose commitment to live healthier. You respond, yes, and plan to join the local gym and exercise more, and put a reminder in your phone to subscribe to a gym. This year will be different and I will be healthier, you say, and you are not alone in doing so. Many individuals will join a gym in January, in pursuit of New Year’s resolutions to be healthier. But will this year actually be different?

New Year’s is a popular time to “turn over a new leaf”, where individuals intend to make changes to their routine to pursue long-term goals. Our understanding of New Year’s resolutions has been benefited from behavioral science insights, as prominent research has highlighted two main reasons why this phenomenon occurs: 1) long-term goal pursuit is challenging, and 2) New Year’s presents a unique motivational opportunity. However, maintaining New Year’s resolutions is another challenge in itself. To preserve this motivation, behavioral researchers have explored methods to help individuals persevere in pursuit of their long-term goals.

Challenge of Long-Term Goal Pursuit

The first component defining New Year’s resolutions involves the challenging aspect of long-term goal pursuit. While we know how important a goal may be, the benefits of achieving the goal are difficult to measure in the moment. If a person hopes to live healthier, the individual may recognize the importance of the goal, but may be unable to realize the outcome of being healthier, since the rewards are distant and abstract.

Therefore, the most salient rewards are those that occur instantaneously. [5] Consider running on the treadmill: running is good for you in the long term, but is dull and effortful in the moment. The dull momentary experience of running becomes our focus when the time for decision-making arrives. Immediate rewards are the driving force behind our decisions, and in this case, the healthier option loses. [5] The short-term barriers of goal pursuit prevent us from achieving our healthy objectives. But why does the New Year see a huge increase in the creation of these objectives?

New Year’s as a Temporal Boundary

New Year’s is a unique time for goal pursuit due to the societal constructs of temporal boundaries. Temporal boundaries, also known as calendar landmarks, mark a new period of time: the start of a new week, a new month, or a new year. [3] Temporal boundaries not only indicate a transition to a new period of time, but also a transitional opportunity for individuals. We perceive the self to have different qualities before and after a temporal boundary. Our current self is stuck in our current routine, like couch-sitting instead of running, but the self after the temporal boundary is open to routine changes, due to the perception of a new chapter of time. [3] “A fresh start” in the calendar presents a fresh start for behavior.

We perceive our future self to be more capable of pursuing challenging goals, especially following a temporal boundary. [1] The future self is closer to our ideal self, in that an individual will be closer to reaching a long-term goal, and further from past mistakes. Therefore, once we pass the temporal boundary, we believe we will be more capable of pursuing challenging goals. [1]

Additionally, since long-term goals have future rewards, the future self will be the beneficiary. If in 2018, a person exercises every day, the same individual will experience a healthier life outcome in 2018. The future self can both pursue the goal more effectively, and will be rewarded for taking action. As we think of the changes we hope to make, 2018 sounds like the perfect time to make those changes.

Behavioral scientists found that in an experimental setting, individuals were more motivated to pursue long term goals at temporal boundaries. Individuals’ interest in dieting increased by 14.4% at the start of a new week, by 3.7% at the start of a new month, and by 82.1% at the start of a new calendar year. [1] Similarly, participants’ probability of visiting a gym increased by 33.4% at the beginning of the week, by 14.4% at the start of a new month, and by 11.6% at the start of a new year. [1]

So, what happens when we pass the temporal boundary? Suddenly, it’s 2018, and the couch is so inviting. After the New Year, we are still the same self, and face the same barriers to the pursuit of our goals as before. The challenging aspects of pursuing long-term goals are vivid, while the rewards are distant and obtuse. As a result, all of those January members at the gym lose their New Year’s resolution motivation.

How to Maintain Motivation

What can we do to persist through these barriers? There are two main strategies to consider: 1) introduce positive external factors to the challenging goal, or 2) set detailed implementation intentions. Both of these strategies produce positive results in the research lab, and are likely to benefit your life, too.

The first strategy is to introduce positive external factors. If the main deterrent to pursuing long-term goals is the discontent we face in the short-term, why not modify the activities? Behavioral researchers explore new ways to make unfavorable tasks more enjoyable. A group of researchers from the University of Chicago introduced some techniques to improve classroom activities for students. These techniques involved introducing positive external factors, like snacks or music, for the students to enjoy during classroom activities.

The researchers found that students spent more time on classroom activities when the music or snacks were available, compared to control groups without the external factors. The same effect happened with exercise – individuals spent more time on their workout routine with positive external factors, such as music, than the control group. [5] Individuals can modify activities to make the immediate aspects positive and enticing. Introducing positive external factors to the activities involved in long-term goal pursuit can help students or aspiring runners engage in the necessary activities in the short term with positive spins, generating better long-term outcomes.

Taking it Step by Step

The second strategy is to form detailed implementation intentions. Striving for a long-term goal is important, but often has weak motivational forces in the moment. Instead, outlining the specific steps necessary to reach the goal throughout the goal pursuit process can help motivate every action. Implementation intentions specify the details of goal pursuit, such as the place, the time, and the instructions on how to perform the action. [3] For exercise goals, aspirational runners can plan their running time, route, and intensity in advance, which provides clear guidance for action execution beforehand.


The AI Governance Challenge

Researchers found that implementation intentions supported long-term goal pursuit of exercise goals better than other motivational tactics. In an experiment conducted on undergraduate students in the United Kingdom, the students that set implementation intentions were significantly more likely to execute their exercise intentions across three different time points, compared with the control group, and with another motivational experimental condition. [4] Students in this condition were unique from the other two conditions because they completed the following statement:

“During next week I will partake in at least 20 minutes of vigorous exercise on (day or days) _______ at ______(time of day) at/or in (place)______.”

All participants in the implementation intentions condition exercised at the location they planned. 97 percent of students in this condition exercised at the intended time, and 88 percent exercised on the planned day. [4] The researchers confirmed their hypothesis that individuals who form implementation intentions about the time, date, and location of exercise, will execute their actions at the specified times. Implementation intentions effectively produced lasting behavior change.

If you’re motivated to make behavior change this New Year’s, pursue it. Setting goals and making plans to achieve those goals are important skills for creating brighter futures. Keep in mind, however, that the motivation to pursue your goals may diminish in 2018. Pair your action plan with fun external factors, or create detailed plans for executing behavior change. Behavioral science research suggests that through special attention to the immediate aspects of goal pursuit, behavior change is possible. If you strategize your goals, this year really will be different.

The Devil You (Expect to) Know: Political Polarization

In his final speech in office, President Obama remarked on the growing level of extreme polarization, particularly in online interactions, by telling those in attendance “[i]f you’re tired of arguing with strangers on the internet, try to talk with one in real life.” As the rise of social media has highlighted (if not exacerbated) political divides in modern society — enabling us to seek echo chambers among those with whom we agree, and to antagonize our adversaries from the relative security of a world mediated by screens — the President’s advice seemed to point to some deeper truth of human interaction: it is harder to demonize someone we know.

“[i]f you’re tired of arguing with strangers on the internet, try to talk with one in real life.”

 The Effects of Mere-Exposure

Indeed, a host of psychological research supports this intuitive claim. Beginning with Zajonc’s[1] (1968) demonstration of the attitudinal effect of “mere exposure” — that simply becoming familiar with some stimulus (a person, animal, object, etc.) is enough to positively influence our attitudes toward that thing — researchers have found the effect present in as diverse contexts as racial attitudes (Zebrowitz, White, & Wieneke, 2008) and tastes for tropical juice (Pliner, 1982). There are two main explanations for how this works: misattribution (i.e., subjects mistake the cognitive and perceptual fluency that comes with exposure to a stimulus for liking that stimulus) and uncertainty reduction (i.e., not knowing about a stimulus causes cognitive stress, which familiarity with that stimulus naturally moderates) (c.f. Lee, 2001). Weighing the evidence for each, Lee (2001) contends that, while misattribution may explain changes in cognitive judgments, only uncertainty reduction can explain how exposure impacts our affective judgments. That is, how much we like or dislike something or someone. Of course, a change in how we affectively judge our political opponents is what President Obama was suggesting could improve discourse. The more we view our adversaries with negative affect, such as contempt, the less likely we are to consider their perspectives.

Berger and Calabrese (1975) provide further axioms for understanding how uncertainty reduction affects interpersonal interactions. Notably, they suggest “similarities between persons reduce uncertainty” (axiom 6) and “decreases in uncertainty level produce increases in liking” (axiom 7). If we meet our adversaries in person, our uncertainty about them decreases — and as this happens, so too should our disdain dissipate. Anecdotally, this phenomenon should resonate with many of us: are not a great deal of our social interactions based on proximity and familiarity?

Short of arranging in-person meetings for all social media squabblers, what can this literature tell us about engendering understanding in political discourse? In her authoritative overview of motivated reasoning, Kunda (1990) further highlights how even anticipated interactions improve our perception of others. In one example, participants in a study observed a taped discussion between three individuals, and were told they would go on a date with one of the three. When asked to rate those three individuals, participants projected onto their hypothetical dates higher scores for personality and perceived likability, and expressed a greater degree of confidence in their projections (Berscheid, Graziano, Monson, & Dermer, 1976; c.f. Kunda, 1990). As Kunda writes, participants’ desire to expect a pleasant interaction positively biased their perceptions of the individuals.

Imagined Contact Hypothesis

More recently, a robust strain of literature has emerged to examine this effect of imagined contact. Making use of insights from mere exposure and contact hypothesis, Crisp and Turner (2009) set out to examine how even a hypothetical interaction with an individual affects one’s perception of her. There are two critical insights from this research: first, that even imagined contact leads to an increase in positive affect for the hypothesized individual (Crisp & Turner, 2009); and second, perhaps even more critically, that this effect is greater when the imagined contact is with an out-group member (Stathi & Crisp, 2008). That is, the benefits of hypothetical interaction are even greater when it is with a person outside of one’s familiar social “in-group” (i.e., age, race, gender, sexual orientation, political party, etc.). Moreover, this effect persisted regardless of whether the interaction had positive or neutral valence — so that any anticipated exposure had a positive impact on behavior.

How We View Our Political Rivals

In the context of uncertainty reduction, this all seems to make sense. We already know that people in our in-group (e.g., members of our political party) share at least some of our own characteristics or beliefs. Thus, imagining contact with them should do little to change our perception. Conversely, people in out-groups inherently have some element of foreignness (this is why they’re not in our in-group), which makes us uncertain about them. By exposing ourselves to them (whether in person or in our imaginations), we are forced to face this unfamiliarity, and in doing so we naturally look for points of similarity. This in turn reduces our perceived uncertainty about them, and causes us to like them more.


The AI Governance Challenge

The mechanism by which this change happens is as follows: with no expectation to meet another individual, I will simply rely on my predominant stereotype in making my judgment of her; however, if I expect to meet that individual — and thus have the directional goal of a pleasant interaction — I will draw more selectively from my cognitive heuristics and memory to construct a more likable picture of the individual.

What all of this suggests is that we desperately need to humanize our politics. The insight that something so small as expecting to meet an individual can meaningfully change our beliefs about and behavior towards them should be a powerful tool for social interventions aimed at reconciliation. Recognizing the person on the other end of that Twitter argument as a human being — with thoughts and feelings, and perhaps even friends and a family — could be a powerful antidote to widespread demonization of our political rivals.