Stir your inspiration and make your research useful

How to choose a research topic

As you arrive at this website, you’re probably thinking “What is the best way to choose my research topic? What I should be guided by when choosing?” Here are some suggestions and ideas:

When choosing a topic there are two sources of information you can draw from - internal (your own impressions and feelings) and external (what other people say and suggest). First, let’s discuss the internal.

Useful considerations for a topic choice

We think it is good to select a research topic based on your interests. This is because being genuinely interested in a topic will motivate you and will help you develop your research taste, which might be a powerful intellectual tool for orienting in complex and not yet defined waters. However, there are two other considerations we think are important.

First: How valuable would an answer to your research question be? What would the consequences be of having that question answered? Would it have any effect at all? Are there any questions that may elicit more valuable answers? Is there a question that could solve more important, larger, and more neglected problems? How much would answering this question improve the world? By “improving the world” here we don’t necessarily mean doing something applied - you can improve the world by testing interventions and finding answers to very applied questions, but you can also improve the world by improving the theoretical understanding of fundamental parts of some problem. Importantly, however, the extent to which this increased understanding is actually valuable may vary significantly depending on the topic. Gaining an internal sense of how valuable the answers to various questions would be is especially important when you want to make progress on open, not yet defined questions and in early-stage fields (which are often very interesting and provide opportunities for greater progress and discoveries).

venn_theses

The second factor is your personal tractability - do you feel that you would be able to make progress on this topic? Even if it is in general possible to make progress on it, are you a good fit to do so? It’s often good to spend about 10 % of your time testing your fit with a project, rather than fully committing to something right away.

These factors are also important to keep in mind when assessing external sources of information. Look at what topics the most prominent researchers are interested in, and listen to what other people and communities think is valuable to work on—paying close attention to their reasoning behind it. Experienced researchers and other well-informed individuals can also be useful resources when determining tractability.

How to compare information from external vs internal sources

Generally, we think it’s good to start with and get inspired by external sources of information, but ultimately let the internal sources have the last word. For example, we think it might be good for you to choose a field/general topic that others have a good reason to say is valuable and that may improve the world more than other fields. Then you can let yourself be guided by your internal impressions on what specifically within this broader domain/general topic feels interesting and tractable.

To give you a head start, we have put together a list of general topics/domains that seem to be very valuable to make progress on from an impartial welfarist perspective (i.e., promoting wellbeing, with every entity’s wellbeing counting equally). Dive in and get inspired!

Explore and get interested

There are several paths we think might be especially high impact in making the world a better place. If you get interested in any of these topics, we can connect you with researchers working in these fields or provide other types of support (scroll down for more info). If you would appreciate more tailored advice, you can try our thesis topic coaching.


Governance of artificial intelligence
#international relations #comparative politics #political theory

Why is this important:
Developments in artificial intelligence algorithms will likely have a large influence on many domains of our societies and lives. Research into how we should govern the creation and deployment of such developments will help us capture the benefits and mitigate the risks stemming from abrupt changes. To learn more, read this guide, watch this and this talk, listen to this, this and this podcast.

How to tackle this:
Generally, questions like “How can we avoid a dangerous arms race to develop powerful AI systems? How can the benefits of advanced AI systems be widely distributed? How open should AI research be?”

There are a lot of questions you can draw from this research agenda, for example, what will the impact of exacerbated inequality and job displacement on trends such as liberalism, democracy, and globalization be; who should be leading on AI governance; how substantial of an advantage does China have, as compared with other advanced developed (mostly liberal democratic) countries, in its ability to channel its large economy, collect and share citizen data, and exclude competitors, etc...

You can also see this guide featuring some additional questions and check some of our finished theses on the topic.

Who are the people already working on this:
Prof Allan Dafoe, Governance of AI team at the Future of Humanity Institute, Center for Security and Emerging Technology, AI Impacts


Improving institutional decision making
#comparative politics #public policy

Why is this important:
Large institutions in our societies, most notably governments, have vast resources and power to change the world both, for the better or for the worse. Yet decision processes used in these institutions are often not optimal or rational. If we could improve these processes even by a little, the expected positive value that would be created is very large. However, we often lack knowledge of how to go about improving these processes. Read more here or watch this talk.

How to tackle this:
Generally, looking into decision making in the group settings would be one option. Specifically, looking into how various methods (like calibration training or structured analytic techniques) improve decisions in real-life scenarios and settings could be valuable. Also, developing better methods for evaluating the quality of arguments when there is no "correct" answer (example 1, example 2, example 3) would be potentially impactful. Another option could be looking into how policy-makers individually work with evidence and update their beliefs (see this paper and this blog post).
Here is a possible framework on how to think about questions related to improving institutional decision making.

Alternatively, looking into case studies on how better practices have been implemented in organisations historically, historical analysis on what happened badly in the event of some big important decisions (for example, the decision to invade Iraq) - e.g. what were the system errors and what were the individual ones; case studies on ballot initiatives on topics related to animal welfare, foreign aid/global development or policies helping future generations/long term future might also bring more understanding of how to contribute to improving institutional decision making on important issues.

Who are the people already working on this:
Philip Tetlock, Eva Vivalt, Wątróbski et al., 2019



Pandemic preparedness

Why is this important:
Large pandemic outbreaks (like recent COVID-19) has historically caused enormous losses in lives (e.g. Spanish flu is estimated to had wiped almost 1 % or possibly as much as 5.4 % of the global population). If we were to tackle more severe natural pandemics or even anthropogenic pandemic caused by engineered pathogens, we might risk significant disruption to global civilisation and the future of humanity’s progress. Working on preventing these scenarios and mitigating them if they come to happen might, therefore, be very valuable. Learn more about the general case for working on biosecurity and global catastrophic biological risks in this problem profile, in this podcast, this podcast and this conference talk.

How to tackle this:
Currently, almost no countries in the world have an action plan describing what to do when such a disruptive event such as pandemics arrives. This leaves the major decision-makers with little to hold onto when it actually happens and they often have to make large and important decisions in an emergency mindset. Trying to draft some plan involving multiple scenarios of how the infection might unfold and suggesting best practices for various timelines would make it easier for decision-makers to identify and implement the best policies when the next pandemic happens.
It would be good to have this plan for each country, as well as a global-level plan.

There are several calculations that this plan would require and some of them don’t seem to exist yet. For example, it would be useful to have an estimate of how many people would be required to keep the essential services (like electricity, water and food supply, hospitals, etc..) running in the event of very severe pandemics requiring very strict lockdown and isolation of as many people as possible.

Another path could be to research how to change the mindset in countries around the world to allow for and incentivise transparency about how many cases of infections they have, when exactly did the outbreak happen, etc… and also establish a data-sharing norm. There are many incentives pushing in the other direction (i.e. secrecy about these types of data) which blocks other countries to effectively prepare for the pandemics or collaborate on stopping the pandemics in the first place.



Other directions that seem promising:

  • International cooperation on reducing global catastrophic risks (especially risks stemming from emerging technologies like artificial intelligence or advanced biotechnology) - example topic 1, 2, see also finished theses on this topic

  • Governance of surveillance (this paper by Nick Bostrom) (conference talk)- surveillance will likely increase in future and we want to make sure it’s governed in a way that produces good outcomes rather than bad (e.g. totalitarian regime)

  • Improving voting systems (cause profile) (podcast) - e.g.case studies of attempts at creating voting reforms -> check also this and this suggestion on how to generally improve democracy

  • Relations with China (problem profile)

If you get interested in any of these topics, let us know. We can:

  • Connect you with researchers working in these fields who can provide feedback on your ideas
  • Help you develop more specific topic ideas
  • Connect you with other students working on the same questions
  • Help you with publishing your thesis

This service is free and paid for by grants from charitable foundations. There are no terms and conditions connected with this service. We only want to help talented students have more impact with their research and support research on the most important problems.