If we can’t save everyone, who do we save?
In the second episode of More Than Good, Effective Altruism Australia’s podcast, we explored one of the hardest questions in philanthropy: when we cannot help everyone, how should we decide who – or what – to help?
We were joined by three guests who brought philosophy, practical experience and plenty of humour to the discussion:
The guests
- Ray O’Leary, writer, comedian and media personality.
- Bridget Loughhead, Head of Community Programs at Effective Altruism Australia.
- Grace Adams, CEO of Effective Altruism Australia
The conversation moved from global health and animal welfare to nuclear winter, the trolley problem and AI-run hospitals. Beneath the hypotheticals was a serious idea: good intentions matter, but when lives and limited resources are at stake, we also need evidence, careful reasoning and the humility to admit when the answer is not obvious.
You can find Episode 2 on:
YouTube
Spotify
Apple Podcasts
If you could only fund one cause
The episode opened with an uncomfortable constraint: if you could fund only one cause for the next 20 years, what would it be?
Ray initially proposed research into which cause would be best to fund – a characteristically effective altruist answer – before suggesting public health interventions in lower-income countries. Bridget added that the starting point should be identifying problems that are highly neglected, likely to remain important and capable of producing meaningful benefits with additional funding.
This introduced one of the central principles of effective altruism: a cause is not necessarily a strong funding opportunity simply because it is visible or emotionally compelling. Its scale, neglectedness and tractability also matter. A problem affecting many people may receive very little attention, while a popular cause may already attract substantial resources.
Sometimes the greatest value comes not from directing enormous amounts of money towards one issue, but from ensuring that a neglected risk receives some serious attention at all.
Bridget used the example of the Alliance to Feed the Earth in Disasters (ALLFED), which develops plans for feeding the world following catastrophes such as a nuclear winter, supervolcanic eruption or other event that severely disrupts conventional agriculture. It may not be a cause most people encounter in daily life, but moving it from almost no preparation to credible planning could be enormously valuable.
Is pessimism stopping us from acting?
Ray observed that negativity is often easier to make funny- and easier to believe. News and social media can make the world appear to be deteriorating constantly, even though many measures of human wellbeing have improved dramatically over time.
Bridget noted that positive claims are often treated with suspicion, as though anyone saying the world is improving must have an agenda. Negative framing, by contrast, can feel more credible and attracts attention more readily.
But constant pessimism carries another risk: if the world feels irredeemably bad, inaction begins to look reasonable. Believing that nothing can be changed can absolve us of the responsibility to try.
Grace offered a radically different picture. If most people donated 10% of their income, she argued, the world could look profoundly different. There could be enough funding to make enormous progress on extreme poverty, expand valuable research and reduce major global risks. Giving might also make people happier and encourage a more cooperative culture.
The point was not that optimism requires ignoring suffering. It was that accurately understanding both the problems and our capacity to address them gives us a reason to act.
Humans, animals & the limits of intuition
Amy then posed a difficult trade-off: would you rather reduce extreme suffering for millions of animals, or moderately improve the lives of thousands of humans?
Ray chose the humans, while acknowledging the difficulty of the comparison and his own commitment to veganism. Bridget resisted answering without more information. How severe is the suffering? How substantial is the human benefit? How confident are we that either intervention will work?
Her desire for a spreadsheet became a running joke, but it reflected a serious principle. Our first emotional response is not always a reliable guide when comparing very different kinds of harm. Humans are naturally biased towards members of our own species, while uncertainty about animal consciousness and moral weight makes the calculation harder still.
Grace later identified animal welfare as an area that remains dramatically underfunded relative to the scale of suffering. She also highlighted childhood lead exposure and pandemic prevention. Around one in three children globally are estimated to have elevated levels of lead in their blood, with consequences that may continue throughout their lives. Meanwhile, much of the urgency around pandemic preparedness that followed COVID-19 has dissipated, despite the possibility of a future outbreak being far more destructive.
The episode did not reduce these issues to a single answer. Instead, it showed why comparisons matter. When resources are limited, refusing to compare does not eliminate a trade-off – it merely leaves the decision to habit, popularity or chance.
Do we prefer causes that make us heroic?
Bridget agreed that people often prefer forms of helping where the impact is personal and visible. Her years as a Scout leader gave her the immediate reward of seeing children enjoy the activities she helped provide. That experience was valuable, but she stressed that feeling close to an outcome is not the same as creating the greatest possible benefit.
Some of the most effective actions are distant, technical or unglamorous. A donor may never meet the child protected from malaria. A researcher working on pandemic prevention may never know which catastrophe their work helped avert. A person delivering vaccines may still wake up on Monday morning wishing they did not have to go to work.
Doing good does not always feel heroic. Sometimes it is repetitive, administrative or invisible.
That does not mean the positive emotions associated with giving are bad. Grace argued that helping others can – and should – feel good. The problem arises only when personal gratification becomes more important than the people or animals being helped.
Bridget added that waiting for an overwhelming feeling of compassion can become another excuse for inaction. People do not need a dramatic calling to make a meaningful contribution. They need a genuine willingness to consider what will help and to act consistently.
Are the most popular causes most important?
The panel’s answer was no.
Grace used cancer research as a difficult example. Cancer is devastating and continued research is important. But when the same resources could prevent children dying from well-understood and treatable diseases at a fraction of the cost, popularity alone cannot determine priority.
As she explained, everyone is someone’s child, and the tragedy of a death does not depend on whether it was caused by a disease that receives extensive attention in a wealthy country or a preventable illness elsewhere.
Bridget raised donations to already well-resourced schools and cultural institutions. Supporting a child’s school or improving one’s local community may be generous and worthwhile, but it can sit close to self-interest.
When weighed against interventions that could prevent severe suffering or save lives, it is reasonable to ask whether all activities described as philanthropy are achieving comparable benefits.
The discussion was not an argument for abandoning every local or personal cause. It was an invitation to distinguish between spending that benefits our own communities and giving intended to help those in greatest need – and to evaluate the latter by what it actually achieves.
Should an AI kill one person to save five?
Amy’s hypothetical imagined a perfectly human-aligned AI running a hospital. Five patients need organ transplants to survive. A healthy person arrives for a flu shot. Should the AI kill that person and use their organs to save the five?
The arithmetic suggests a net gain of four lives. Almost every other moral instinct says no.
Bridget explained that moral uncertainty gives us strong reasons to respect constraints such as the prohibition against killing an innocent person. Even if consequentialism offers useful guidance in many situations, violating a fundamental rule creates an enormous risk of moral error.
There are also practical consequences. A hospital that might harvest healthy patients’ organs would destroy public trust. People would avoid medical care and vaccination, potentially causing far more deaths. Rules against killing, lying, stealing and coercion often produce better outcomes precisely because cooperation depends on people being able to trust one another.
Ray connected this to rule utilitarianism: even where breaking a rule could appear beneficial in one isolated case, following the rule reliably may create far better consequences across society.
Grace offered a useful rule of thumb: move much of the way towards utilitarian reasoning, then stop before it overrides every other moral consideration. Effective altruism is not a demand to pursue numerical gains at any cost. It is an attempt to use evidence and reason to help others while remaining alert to uncertainty, rights and the systems that allow society to function.
Where the maths leads somewhere disturbing
The episode then moved into one of population ethics’ most famous problems. Imagine a small world in which ten million people live exceptionally rich and joyful lives. Now compare it with a world containing trillions of people whose lives are only barely worth living. If the second world contains more total wellbeing, should a purely mathematical approach prefer it?
The panel found the conclusion deeply uncomfortable. Grace felt the numbers mattered but struggled with the intuition that a vast population living mediocre lives should be preferable. Bridget noted that every moral framework produces strange answers in extreme edge cases. Rejecting a useful ethical approach solely because of one bizarre hypothetical may leave us with no practical guidance at all.
She drew on the idea of moral uncertainty: we may have strong reasons to value total wellbeing while remaining unsure whether happiness is the only thing that matters. Art, excellence, freedom, relationships and other values may also carry intrinsic importance. When we do not know which moral theory is correct, a mixed approach may be wiser than complete confidence in any single framework.
The biggest mistakes in giving
Ray suggested that much charitable giving is reflexive. Someone approaches us in the street, a story catches our attention, or a cause appears at the right emotional moment, and we donate without comparing what our money might accomplish elsewhere.
Effective altruism tries to preserve the generous impulse while asking a further question: how can that desire to help benefit as many people – or sentient beings – as possible?
Bridget proposed an even more basic mistake: not giving at all. Many people want to contribute to a better world but never incorporate giving into their plans. Establishing the habit can be the first step. Once someone has committed resources, they are often more motivated to investigate where those resources will do the most good.
The episode’s conclusion was not that every ethical dilemma can be solved by a spreadsheet. It was that careful comparison is better than leaving life-and-death decisions to familiarity, popularity or impulse. We may not be able to save everyone, but we can take seriously the responsibility to help as many as we reasonably can, and to keep improving the way we decide.
More Than Good is a podcast from Effective Altruism Australia.
You can find Episode 2 on:
YouTube
Spotify
Apple Podcasts
