The introductory program in effective altruism

Scope sensitivity means noticing that the number of beings affected carries moral weight. Saving a thousand lives matters more than saving ten.

This does not mean that numbers are all that count. Also important are, among other things, the depth of the improvement, the durability of the effect, the costs, the probability of success, and possible side effects.

Because our time, attention, and resources are limited, we cannot support all worthwhile causes at once. In practice this means having to choose between different ways of making the world better.

Effective altruism encourages making such choices deliberately: asking where an extra hour of work, an extra zloty, or a particular skill could produce the greatest expected effect.

Scout mindset is an orientation toward discovering the truth rather than defending one's existing beliefs. In the context of helping, it means a willingness to ask difficult questions, look for evidence, compare options, and change one's mind in the face of good arguments.

This does not always come naturally. We often tend to defend our existing beliefs, seek out information that confirms our intuitions, or avoid inconvenient conclusions. If we want to help others more effectively, however, it is worth developing the ability to think clearly and to honestly test our own assumptions.

Moral impartiality means striving to take into account the good of everyone, regardless of where and when they live, how similar they are to us, and how easily we can identify with them.

It does not mean that we should stop caring about those close to us or about our local community. Rather, it means being ready to ask: whose needs are greatest, where is help most neglected, and where could our actions make the biggest difference?

We will return to this concept in later sessions – first in the context of global differences in needs and opportunities to help, then with regard to non-human animals and future generations.

Some of the ways of helping available to us can do many times more good than others – for a similar outlay of time, money, or attention. These differences are often far greater than we intuitively assume.

For that reason it is worth asking not only whether a given action helps, but also: how much it helps, whom it helps, at what cost, and whether the same resources could achieve a greater effect elsewhere.

Cost-effectiveness helps us compare how much good can be achieved with a given quantity of resources – for example, a single donation, an hour of work, or a social program.

It is not a perfect tool. Not everything important is easy to measure, and numbers can create a misleading sense of certainty. Even so, simple estimates often help us make better decisions than intuition alone.

In many decisions we do not know for certain what is true or what the consequences of our actions will be. Expected value is a way of thinking that helps us make decisions under uncertainty: we take into account the possible outcomes, their scale, and their probability.

This tool is particularly useful when comparing actions with different levels of risk – for example, well-researched but moderately effective interventions versus more uncertain ones that are potentially far more impactful.

The ITN framework helps compare global problems in terms of their expected impact. Its name comes from three criteria:

• importance – how big the problem is and how much good could come from solving it;
• tractability – how much progress additional resources could enable;
• neglectedness – how little attention, money, and work is already devoted to the problem.

The ITN framework does not give an automatic answer as to what should be done. It does, however, help structure our thinking and notice which problems may be especially promising from the perspective of additional action.

In this session we return to impartiality in a new context. In Session 1 we asked whether geographical distance should affect how much someone's well-being matters. Now we ask whether species, in itself, should decide whose suffering we take seriously.

Impartiality does not automatically mean that all beings have exactly the same interests, or that every difference between them is morally irrelevant. Rather, it means that differences such as where someone lives, how long they live, or what species they belong to should influence our decisions only when they really are morally relevant.

We are often unsure how much a given action will help or what its side effects will be. In such situations we can weigh the possible outcomes, their scale, and their probability, and then choose the action that looks best in light of the available information.

Thinking in terms of expected value can be helpful when comparing interventions with different levels of certainty. It remains, however, only a tool: it requires care, sound assumptions, and an awareness that not everything morally important can be easily captured in numbers.

Throughout history, many widely accepted beliefs have turned out to be wrong – including on matters of great moral significance, such as slavery. It is therefore worth staying ready to consider ideas that may at first seem strange, counterintuitive, or unpopular.

Openness does not, however, mean uncritically accepting every unusual claim. A good attitude combines curiosity, epistemic humility, and critical thinking: we ask what the arguments are, what the evidence is, what the risks of error are, and how we can act responsibly toward others.

Animal welfare concerns the quality of animals' experiences: suffering, pain, stress, frustration, the ability to engage in natural behaviors, and, more broadly, whether their lives go well or badly from their own perspective.

From the perspective of effective altruism, this topic may be especially important because a very large number of animals live in conditions of industrial farming, and their interests are often poorly represented in public debate, policy, and philanthropy.

Longtermism is the view that positively influencing the very long-term future may be one of the most important moral priorities of our time.

Its basic intuition is this: if future generations will be numerous and their well-being matters morally, then actions that shape the long-term future may carry enormous weight. This applies especially to actions that reduce the risk of catastrophes capable of ending civilization's development or permanently restricting its possibilities.

Longtermism need not imply that present-day problems are unimportant. Rather, it means that in making moral decisions it is worth taking into account also the people and other beings who may exist in the future.

An existential risk is the risk of an event that could lead to human extinction or permanently and drastically curtail humanity's long-term potential.

Examples of such risks include certain scenarios involving advanced artificial intelligence, biological catastrophes, nuclear war, extreme climate change, and other global catastrophes. They differ in probability, scale, degree of neglect, and how easily they can be countered.

In Session 1 we considered impartiality toward people living far away from us. In Session 3 we asked whether our concern should also extend to non-human animals. In this session we examine the third dimension of the same idea: time.

The question is: does the fact that someone will live a hundred, a thousand, or a million years from now, in itself, diminish the moral significance of their well-being? If not, then future generations may deserve more serious weight in our decisions than we usually give them.

This does not yet settle which actions matter most. It does, however, show why longtermism and work on existential risks are taken seriously by part of the effective altruism community.

AI safety encompasses efforts to ensure that increasingly powerful AI systems are designed, deployed, and governed in ways consistent with human values and resilient to serious errors, misuse, and unintended consequences.

From the perspective of existential risks, particularly important questions include whether advanced AI systems could become difficult to control, whether they could accelerate the development of other risks, and whether social institutions can keep pace with the speed at which this technology is developing.

Biosecurity concerns limiting the risk of serious threats arising from infectious diseases, natural pathogens, laboratory accidents, or the deliberate harmful use of biology.

What matters in this area are both technical and institutional measures: better disease surveillance, faster detection of outbreaks, safe laboratory practices, responsible management of research, and the preparedness of public health systems.

Longtermism and work on existential risks involve great uncertainty. It is difficult to precisely estimate the probability of rare, extreme events, and harder still to assess the effectiveness of actions intended to prevent them.

It is therefore worth using the tools learned earlier: the ITN framework, expected value, scope sensitivity, awareness of trade-offs, and scout mindset. At the same time, we should remain cautious about overly confident conclusions based on speculative assumptions.

For many people, the choice of career may be one of the most important decisions about their impact on the world. Work takes up a large part of life, develops skills, builds networks, and can directly or indirectly affect important global problems.

From the perspective of effective altruism, it is worth asking not only "what work interests me?" but also: where could my skills be especially useful, which problems are important and neglected, how can I build career capital, and which paths offer the greatest chance of positive impact.

Donations are one of the simplest ways of turning resources into real help. Their impact can vary greatly depending on the cause and the organization they go to.

Giving pledges, such as the commitments proposed by Giving What We Can, can help make giving serious and consistent. They need not, however, be the first or only step. For some people a good starting point may be a trial pledge, a small regular donation, or a one-off gift to an organization recommended on the basis of the available evidence.

Effective altruism is not a closed set of answers but a community and a way of thinking that keeps developing. After the program ends, you can continue exploring different areas: global health, animal welfare, AI safety, biosecurity, public policy, metascience, community building, grantmaking, communication, or your own projects.

Further exploration might include reading the EA Forum, talking with people from the local community, attending events, career consultations, volunteering, small project experiments, or an in-depth analysis of a particular problem area.

It is worth taking giving seriously, but this does not mean it should replace all other values. People have more than one goal: relationships, health, personal development, stability, rest, creativity, family, community, and local commitments.

A mature approach to effective altruism should take into account both moral ambition and the sustainability of effort. In practice, it is often better to build a realistic, long-term commitment than to make decisions under pressure, guilt, or sudden enthusiasm.