Future Risks for Sentient Beings: An Introduction

Original text Luciano Carlos Cunha · Volume IX, p. 10 of the PDF edition · English translation

1.1. The past, the present, and the future

The emergence of factory farming in the 1930s and its subsequent large-scale application resulted in an enormous multiplication of two things: (1) the number of animals killed and subjected to living a life of suffering; (2) the levels of suffering for each animal. In short, factory farming multiplied the harms humans cause to other animals on a scale never seen before.

At the beginning of the twentieth century, animal advocates probably would not have imagined that factory farming would emerge just a few decades later. However, at that time there were already clues that the technology then emerging would result in factory farming.

Now consider two thought experiments¹.

(1) You, going from the present to the past

Imagine that you could go back in time and warn those people that something terrible was about to emerge. Perhaps there would be time to prevent factory farming from emerging (or at least to reduce its negative impact). Of course, it might be too late to try anything. But imagine that your time machine allowed you to go back to well before the beginning of the twentieth century. It seems that the earlier you managed to go back, the greater your chances of preventing the catastrophe (or at least of minimizing its negative impact), since the people you influenced would have more time to try.

(2) You, coming from the future to the present

Now imagine that you are once again a time traveler, but this time from the future. Imagine that you have come from the year 4025 to warn the people of the present that things are about to emerge

in the future so terrible that they would make even factory farming look like "a tiny grain of sand" by comparison. As in the previous example, there is a reason for you to come back to the present moment. At first you tried going back to times closer to 4025, but you saw that your efforts would be futile: it was necessary to go back to earlier times to allow enough time to prevent the catastrophes. For this reason, you ended up in the present time.

What these thought experiments aim to ask is: with regard to the bad things that could emerge for sentient beings² in the future, are we now perhaps in the same position that animal advocates were in before the emergence of factory farming? As we will see below, this is what the concept of s-risks is about: suffering on a gigantic scale that could arise in the future, and what to do to prevent it.

1.2. What are s-risks?

At present, a very large number of nonhuman animals suffer intensely, and in many cases in ways that did not exist in the past. Likewise, in the future new ways of harming sentient beings on a massive scale may emerge, on a scale vastly greater than anything that has existed so far.

One concept used to address this problem is that of risks of future suffering (usually abbreviated as s-risks). This concept concerns the possibility that suffering will be massively increased in the future³, and it can be defined as follows:

• S-risks: risks that events will occur in the future that result in suffering of gigantic magnitude⁴, exceeding the suffering that has existed on Earth so far.

The concept of s-risks is centered on suffering, but there could be an analogous concept to address all the forms of harm that animals could endure, and not only suffering

(such as, for example, the harm of death). In that case, there could be a concept such as h-risks (risks of future harm). However, in this book we will speak of s-risks because that is the concept that has been widely used in the debate on these questions, even though, technically speaking, h-risks would be a more appropriate term, since it is more comprehensive.

1.3. Differences between s-risks and x-risks

S-risks are a different concept from x-risks (existential risks). Both s-risks and x-risks are about the long term and astronomical impacts. But whereas x-risks are about risks of the annihilation of humanity (or at least of a drastic reduction in its potential), s-risks are about scenarios involving astronomical suffering (and it is therefore a concept that necessarily takes into account all sentient beings).

Certain s-risks may also be x-risks, but there are s-risks that are not x-risks. For example, there are forms of suffering inflicted on nonhuman sentient beings that do not harm humans, and that may even promote certain human interests (the consumption of animals is an example).

1.4. Measuring the seriousness of an s-risk

Every s-risk concerns risks of suffering of astronomical magnitude, but each s-risk has a different degree of seriousness. This is because not only does the severity of each s-risk vary (for example, the number of victims who would potentially be affected and how much each one would suffer), but the probability of each s-risk materializing also varies.

Thus, one way to measure the seriousness of an s-risk is to multiply how bad the outcome would be by its probability of occurring (this would reveal its expected disvalue, or ED)⁵.

This calculation reveals something important about s-risks. Since s-risks are, by definition, many orders of magnitude greater than current suffering (even compared with animal exploitation combined with the suffering of animals in the wild), the ED of each s-risk is enormous, even when the probability of its occurring is small.

1.5. S-risks and new technologies

One factor that can give rise to s-risks is the emergence of new technologies. This risk is great especially if the technology would benefit those who hold it and they have no moral consideration for the beings negatively affected by it. If humanity has advanced technology but moral progress does not occur in parallel, there is a great risk that this will result in suffering of unprecedented magnitude.

As we have seen, one past example of how the emergence of a technology massively multiplied the suffering and deaths of sentient beings was the advent of factory farming (which, in turn, was a result of industrialization). Another example was the development of new weapons technologies.

If human extinction or a collapse of civilization does not occur, technology will probably continue to advance, perhaps making it possible to colonize planets and even billions of galaxies. Thus, technologies with a vastly greater impact may emerge in the future, which also increases the chances that their negative impact will be vastly greater.

For scenarios of astronomical suffering to occur, it is not necessary for the agents to intend to harm the victims: it is enough that they do not care (or care very little) about them, as happens with the act of consuming animal products, for example. New technologies, together with indifference, are already enough to considerably increase the probability that suffering of colossal size will be caused.

Moreover, it may also happen that the agents intend to use the technology to harm the victims. Or, again, suffering on a massive scale may arise accidentally. It is also possible that agents allow or promote the victims' suffering of natural harms when they could have prevented it. In short, there are many ways in which an s-risk could arise. We will see more about this below.

1.6. Classification of s-risks according to how they arise

S-risks can be classified according to various parameters. In this section, we will see how they can be classified according to the way they arise, following the classification given by Baumann (2022, p.13-9). In the next section, we will look at other parameters.

1.6.1. Incidental s-risks

These occur when ways of achieving a certain goal generate a great deal of suffering, but without the suffering being sought for its own sake. The agents may simply be indifferent, or may even prefer an alternative without suffering, but not be willing to pay the costs.

Some examples:

• Suffering as an effect of high productivity. A present-day example is factory farming. In the future, similar scenarios may occur, but on a much larger scale.
• Suffering for the sake of gaining information. A present-day example is animal experimentation. An example in the future would be the creation of virtual simulations with digital sentient beings to gain knowledge about how sentient beings behave.
• Suffering for entertainment. Just as there are bullfights today, new kinds of sentient beings may be created in the future for similar forms of entertainment.

1.6.2. Agential s-risks

These occur when the agent aims to cause the harm. Let us look at some examples:

• Sadists. Future technologies may multiply their potential to cause harm.
• Feelings of hatred. A tribal "us vs. them" mentality can result in hatred against those who belong to a certain ethnicity, religion, political ideology, etc.
• Extreme punishments. Past examples are the forms of punishment that were applied precisely because they caused extreme suffering.

A key factor in the increase of agential s-risks is the presence of malevolent traits (narcissism, psychopathy, Machiavellianism, and sadism) in powerful individuals⁶.

1.6.3. Natural s-risks

These arise without the action of agents. For example, independently of human action, natural processes already harm animals to an extremely high degree, tending to maximize the number of sentient beings who are born to have lives full of suffering⁷. For every animal who manages to survive, literally thousands or even millions are born only to experience suffering and die prematurely. The norm in the wild is for them to be victims of malnutrition, hunger, and thirst, disease, physical injuries, psychological stress, hostile weather events, natural disasters, and interspecific, intraspecific, and sexual conflicts.

These animals are the vast majority of sentient beings on Earth. For example, if we compare the populations of animals exploited by humans and of animals in the wild and draw an analogy with the span of one year, the former would represent only 14 seconds of the year⁸. All the rest (364 days, 23 hours, 59 minutes, and 46 seconds) would be animals in the wild. Natural s-risks are the risks that suffering resulting from natural causes will expand on an astronomical scale.

This suffering is a result of the way natural selection works. For example, the trait of maximizing the number of offspring offers a great advantage in transmitting genetic information: with such gigantic broods (with thousands or even millions of offspring, depending on the species), it is highly likely that at least one or two will survive and reproduce, passing this trait on. However, in stable populations, on average only one offspring survives per adult⁹ (that is, two per brood). All the rest are born almost solely to experience the suffering associated with death.

If this maximization of suffering is a result of natural selection, then, if there are other

planets where sentient life has developed, it is likely that this has also happened there and continues to happen.

There is also a risk that humans will in the future expand this suffering (for example, by spreading life beyond Earth). Although this would already be an incidental s-risk, it would also, indirectly, be a natural s-risk.

1.7. Other classifications

S-risks can be classified according to various other parameters, and not only according to the way they arise. Below are some other possible classifications¹⁰:

• Known and unknown s-risks. These depend on whether or not we are currently able to imagine them (for example, in the Middle Ages no one imagined that the atomic bomb would come about).

Unknown s-risks can also be subclassified according to the way they arise:

• Incidental: when unknown mechanisms will lead to incidental s-risks.
• Agential: agents may come to have reasons in the future to cause harm that we do not yet know about.
• Natural: the universe may already contain much more suffering than we think.

Other parameters according to which s-risks can be classified are:

• According to the type of sentient being affected. For example, whether or not it is human, whether or not it is organic, etc.
• Whether or not they can be influenced. That is, whether or not it is possible to do something to prevent or minimize them. Here, however, it is important to distinguish between: (1) scenarios that, given the laws of physics, it is impossible to influence and (2) scenarios that cannot currently be influenced but may become open to influence in the future.

There may be many other parameters according to which s-risks can be classified. The list above should not be seen as complete.

Below we will look at some concrete examples of s-risks.

1.8. Some examples of s-risks

In the debate about the future, there is a concern with trying to estimate what the main s-risks are and how to prevent them (or at least how to make them occur to a lesser extent). Let us look below at three possible scenarios of massive suffering in the future and the risk factors that could give rise to them. The risk factors will be discussed in more detail in Chapter 8. In Chapter 9 we will look at some strategies for preventing them.

Important: since it is difficult to estimate the probability of the future turning out one way or another, the following examples are well-informed guesses, but they are not necessarily the s-risks most likely to occur. Moreover, all of these examples may still be a small fraction of s-risks, compared with the s-risks still unknown.

S-risks can arise from a wide range of scenarios. Therefore, we must be careful with the availability heuristic, which is the bias that leads us to think that the cases we remember most easily are necessarily the most representative. This bias can lead us to think, mistakenly, that the s-risks we remember most easily will necessarily be the most likely.

With this in mind, here are some examples.

1.8.1. An increase in animal exploitation

There are at least three ways in which the suffering caused by animal exploitation could come to be massively increased:

• Expanding it in scale (that is, affecting a greater number of sentient beings).
• Making it qualitatively worse (that is, increasing the harms suffered by each animal).
• Creating new ways in which it can happen.

One example of an s-risk of this kind that has materialized is the use of insects for consumption, which emerged at the end of the twentieth century and has been expanding in several countries¹¹. Since these animals are very small, producing an equivalent amount of meat requires a vastly greater number of individuals to suffer and die¹². Moreover, invertebrates are typically disregarded even more than other nonhuman animals. This degree of disregard magnifies the negative effect on each invertebrate. For example, the space allotted to insects on farms is extremely small, even smaller, in proportion to their size, than that given to other animals¹³.

1.8.2. The expansion of wild animal suffering

An increase in the suffering of wild animals may occur owing to an increase in the number of animals on Earth (as with rewilding¹⁴ practices, which aim to return ecosystems to their state prior to human presence) or owing to the intentional or accidental spread of life beyond it (as in panspermia¹⁵ and terraforming¹⁶).

These practices may be good from an environmentalist point of view or an anthropocentric one, but they are very negative for the animals affected. As we have seen, natural processes tend to maximize the number of animals who live lives full of suffering and die prematurely. This happens on a gigantic scale, much larger even than that of animal exploitation¹⁷, and it happens to a greater extent the greater the amount of plant biomass available, since this implies an increase in the number of reproductions, which, in turn, is directly linked to the maximization of suffering¹⁸. Given the prevalence of the environmentalist view and the disregard for the good of animals, the expansion of natural territory and of life are usually seen as entirely positive. For the animals affected by such practices, however, their effects are predominantly negative¹⁹.

1.8.3. The creation of new forms of sentience

The creation of new forms of sentience may also give rise to vast amounts of suffering, especially if there is a high probability that these beings will not receive moral consideration. Examples of risks of this kind are the creation of genetically modified animals and, especially, the creation of nonbiological forms of sentience²⁰.

The possibility of nonbiological sentience exists because perhaps sentience appears whenever any physical structure is organized in such a way as to provide the conditions for it to appear, regardless of whether or not the structure is organic. One way to understand this possibility is through the following thought experiment.

Let us imagine that our neurons were gradually replaced by nonorganic neurons. If our sentience did not fade as the neurons were replaced, and if we remained sentient after all the neurons had been replaced, this would suggest that it is possible for nonorganic beings to be sentient in the future, should they come to have a structure organized so as to perform the same functions as a brain.

This possibility is an s-risk because, if these beings come to exist, there is a high probability that they will not receive moral consideration because they are not organic, and will then be harmed in all sorts of ways, just as nonhuman animals are today. The very mistaken idea held by many animal activists, that animals should be respected because they are animals or because they are living beings (and not because they are sentient), is a factor that has the potential to result in disregard for nonorganic sentient beings.

In the future, nonorganic sentience could arise not only through digital modifications of biological bodies, but also, for example, in sentient simulations within a computer program. Such beings, although sentient, might have no faces and might neither move nor scream. So it might be hard to feel empathy for them²¹. If such beings come to exist, humans will probably not care about them, just as happens with animals (or perhaps it will be even worse, since animals are already very similar to humans and yet,

even so, most humans do not care much about them). It may also happen that a digital entity is already sentient at a certain point without our noticing (and that we then cause it suffering without realizing it).

If nonorganic sentience is possible, enormous numbers of digital sentient beings may come to exist in the future, since creating them may involve far lower costs than creating biological sentient beings. Just as anyone can copy computer programs at will today, it is possible that in the future anyone will be able to create digital sentient beings.

Thus, the potential for vast numbers of digital sentient beings, together with a lack of consideration, poses a serious s-risk (exactly as happens in animal agriculture, all the more so since both are combined with economic incentives).

For this reason, it is important to combat not only speciesism, but also substratism. Just as speciesism is discrimination based on the species to which someone belongs, substratism is discrimination based on the type of substrate that makes up a being's body (for example, whether it is organic or nonorganic). Just as anthropocentrism is a form of speciesism (which discriminates against those who do not belong to the human species), carbonism is a form of substratism (which discriminates against those who are not organic).

Substratism, just like speciesism, can produce suffering in astronomical amounts in the future. That is why it is important to emphasize that animals should be considered because they are sentient (and not because they are animals or because they are living beings), and that this implies considering nonorganic sentient beings, should they come to exist.

1.9. Conclusion

The risks that the scenarios mentioned above will materialize are not trivial. In fact, they are very great, since it is easy to ignore such problems and, moreover, they are not being discussed. For this reason, it is extremely necessary to begin a broad debate on s-risks.