Leading Impactful Careers

If you want to make a difference, which issue is best to work on — climate change, education, pandemics, or something else?

People often think that making these kinds of comparisons is near impossible. The most common advice is that you should just work on whatever issue you’re passionate about.

But we believe some global problems are far bigger and more neglected than others, and so which issues you work on will probably be the biggest driver of your impact.

An introduction to comparing global problems

We’d like to see a great many global problems get more attention, but as individuals, the best we can do is identify the biggest gaps in existing efforts and help fill them.

To find these gaps, one starting point is to look for problems that are:

  1. Important: if progress is made, how much social impact would result?
  2. Neglected: how much effort will be invested in this problem by others?
  3. Tractable: how easy is it to make progress per unit of resources?

In this article, we’ll argue that there are huge differences in how important and neglected different issues seem, which don’t seem to be offset by differences in tractability.

This means that by choosing a different issue, you might be able to increase how much impact you have by over 100 times.

Some problems are bigger than others

Climate change is widely considered one of the world’s biggest problems, and we think it’s even bigger than often supposed. While the most likely scenario is several degrees of warming, the uncertainty in climate models means it’s hard to rule out warming over 10°C by 2200.

What’s more, the CO2 we emit today will stay in the atmosphere for tens of thousands of years, impacting our children’s grandchildren and beyond. We think future generations matter, which makes the issue even bigger in scale.

But as we argue in our problem profile on climate change, it looks unlikely that even 13°C of warming would directly cause the extinction of humanity (though it could contribute to making other existential threats worse). As a result, we think there may be issues that are even larger still.

The philosopher Toby Ord has argued that in 1945 humanity entered a new age, which he calls ‘the Precipice.’ On July 16, 1945, humanity detonated the first atomic bomb, which would eventually make it possible — for the first time in history — that a small group of people could destroy most of the world’s cities within hours.

The annual risk of an all-out nuclear exchange is small, but it’s not zero: there is always the chance of an accident or malfunction.

The average of several expert surveys estimated the chance of a US-Russia exchange is 0.4% per year.

The probability of an all-out exchange is lower, but over our lives and the lives of our children, it could still add up to a substantial chance of a catastrophe potentially more devastating than climate change.

Besides killing most people living in urban areas, the resulting fires could lift enough ash into the air to obscure the sun and reduce global temperatures for years, leading to widespread famine through a phenomenon known as ‘nuclear winter.’

Within months, this would not only kill most people alive today, but it could also lead to a collapse of civilisation itself. We think a permanent collapse is very unlikely, but when we consider the scale of the consequences — the loss of all future generations — that risk may be the worst thing about a nuclear conflict.

But the possibility of extreme climate change and nuclear winter are just two examples of a broader trend.

Technology has given this generation unprecedented power to shape history. The consequences of the decisions we make today about nuclear weapons, genetic engineering, artificial intelligence, space settlement, and other emerging technologies could ripple forward for thousands of years, with either enormously positive or enormously negative consequences.

So: some problems are much bigger than others.

Some problems are more neglected than others

In 2016 we argued that a global pandemic posed a significant global risk, but with around $10 billion spent globally on preventing the worst pandemics per year,1 it was more neglected than climate change or international development (which receive hundreds of billions) — which are in turn far more neglected than education and health in rich countries (which receive trillions).

Neglectedness of problem areas

Why work on issues that are comparatively neglected? At least among issues that are roughly similar in importance, it’s usually harder to have a big impact working on more established or popular issues, because there are probably already people working on the most promising interventions. For this reason, if you’re the 100th person working on a problem, your contribution is likely to make a much larger difference than if you’re the 10,000th.

How much larger? In our view, returns to more work diminish relatively quickly, and approximately logarithmically — meaning that it matters a lot how neglected an area is.

From what we’ve seen, some global issues appear to be thousands of times more neglected than others of similar importance — they receive only a tiny fraction of the resources. This implies that if all else is held constant, work in some areas is thousands of times more effective than work in others.

Inspired by this argument, one of our readers, Cassidy, realised that her knowledge as a doctor might be put to even better use in pandemic prevention.

She applied to a master’s in public health programme, and from there was able to get into a biosecurity PhD at Oxford. Since the field of pandemic control was relatively small before COVID-19, she was able to advance quickly. When COVID-19 broke out, she was ready to advise the UK government on policy ideas to control COVID-19, as well as how to prevent the next (potentially much worse) outbreak.

The importance of working on neglected issues means that following your current passions could easily point you in the wrong direction. You’re most likely to stumble across the same issues everyone is already talking about, which will usually be among the least neglected. The best options are probably unconventional.

Which issues might be even more neglected than pandemics?

Three separate surveys of AI researchers at the top machine learning conferences (NeurIPS and ICLM) in 2016, 2019, and 2022 found that researchers believe there’s about a 50% chance that AI systems will exceed human capacities in most jobs around 2060.2 This would be one of the most important events in history.

These same researchers also estimated that if this happened, while the outcome might be ‘extremely good,’ there is also a 5% chance the outcome could be ‘extremely bad’ (e.g. human extinction).3 (That said, it’s unclear how reliable these estimates are — we discuss this more in our profile on preventing an AI-related catastrophe.)

One reason some people are concerned is that it’s unclear whether AI systems will continue to stay aligned with human values as they become more powerful. This could suggest that research into ‘AI alignment’ is a crucial challenge from a long-term perspective.

Since we first wrote about it in 2012, AI alignment has grown into a flourishing field of research in computer science, but it still receives under $100 million in funding per year — about 100 times less than preventing pandemics.

Long-term AI policy — addressing the question of how society and government should handle advancing AI — is even more neglected. (Read more about our case for both.)

How much do problems differ overall?

How important and neglected a problem is multiply together to determine how pressing the issue is overall.

We’ve argued that some issues seem over 100 times bigger than others, and some seem over 100 times more neglected than others.

Since it’s often the most important issues that seem the most neglected, this would suggest that overall you’ll have over 10,000 times as much impact working on some issues rather than others (all else equal).

That said, there are some strong counterarguments to acting as if the spread is this large, which we discuss in our podcast about the key ideas series. Overall, we think it might make sense to act like the differences are more like 1,000-fold.

For instance, based on our view of global priorities, we think by focusing on global poverty rather than a typical social issue in a rich country, you might have 10–100 times as much impact, and then by focusing on existential risk, you might have 10–100 times as much impact again.

You might differ from us in which issues you think are most pressing, but we expect you’ll still conclude there are very large differences.

What does this spread imply?

In an ideal world, there would be far more people working on every important social issue. But each of us only has one career, and we’ll all have far more impact if we focus on the issues that are the most pressing for us to work on.

If it’s possible to have 100 or even 1,000 times as much impact per year by changing the issue we focus on, that’s a huge deal. It would probably be the single biggest thing you could do to increase the impact of your career.

If you don’t have the option of making a big career change right now, there’s a lot you can do to support the most pressing issues no matter your current job, through donations, political engagement and mobilising others. But how about working on these issues directly?

Sometimes it’s relatively easy to support a new issue in your existing role. For instance, if you work in media, you might be able to tilt which issues you cover.

Our readers are sometimes surprised at how their existing skills can be applied to unusual problems like AI alignment.

Brian Tse was working at an investment bank in Hong Kong and didn’t have a background in AI. But he started to learn more about it in his spare time, and using his bilingual background, began translating materials to connect Western and Chinese researchers.

He now runs an independent consulting firm advising organisations in China, the United States, and Europe on the safety and governance of AI and other potentially transformative technologies — a path he also finds much more fulfilling and exciting than banking.

Many people working in government have significant flexibility about which areas of policy they work on. For instance, Clíodhna was working in health policy, but after learning more about AI alignment, she was able to switch to working at the intersection of AI and health.

But even if the switch seems hard, the huge potential gains could mean it’s easily worth it — even if you’d need to retrain, take a more junior role, or test out a role where you’re not sure you’d be a good fit or aren’t sure how to make progress on the issue.

People we advise are often tempted to go for an issue they think is second-tier because it seems easier to enter. But if a top-tier issue might have 10 times the impact, it’s often worth spending some time testing out your fit, even if you’re not sure it’ll work out.

That’s not to downplay the difficulty — orienting your career around a new problem is a big decision. Our main message is that it deserves some very, very serious thought — and much more attention than it normally gets.

Read More: https://80000hours.org/articles/your-choice-of-problem-is-crucial/

Stay in Touch

Join the movement.

Feel free to contact us.

Up