Social Justice and the Curse of Dimensionality

How Effective Altruism Misses the Long Tail

Today marks the start of a significant criminal trial for Sam Bankman-Fried, popularly known as SBF. As the former CEO of the prominent cryptocurrency exchange FTX and co-founder of Alameda Research, SBF was a towering figure in the crypto universe. However, the tides turned when FTX and Alameda Research filed for Chapter 11 bankruptcy in late 2022. Once the 41st richest American and the 60th globally, boasting a staggering net worth of $26 billion, SBF saw his fortune plummet to zero in the wake of the bankruptcy.

The legal woes compounded on December 12, 2022, when authorities arrested SBF in the Bahamas, leading to his extradition to the U.S. An unsealed indictment revealed serious charges, including wire fraud, securities fraud, and money laundering. Though initially released on a $250 million bond and released to his parents’ home, alleged witness tampering led to the revocation of his bail in August 2023, resulting in his return to detention.

Amidst SBF’s legal turmoil, I’m reflecting on his association with the Effective Altruism (EA) movement.

What is Effective Altruism?

Effective Altruism (EA) is championed as a movement grounded in evidence and reason, aimed at maximizing the positive impact one can have on the world. Effective altruists, as followers are termed, are said to choose careers and/or support charities anticipated to render significant societal benefits.

Notable figures associated with EA include moral philosophers and activists like Peter Singer, Toby Ord, William MacAskill, and, not to be forgotten, SBF, who was one of the movement’s significant financial backers. Prominent organizations within the EA sphere include GiveWell, OpenPhilanthropy, Giving What We Can, and 80,000 hours.

In recent years, EA’s interpretation of maximizing positive impact has shifted towards longtermism, a philosophy emphasizing future theoretical scenarios over present, tangible human suffering. This pivot is notably manifested in EA’s significant investment in artificial intelligence. Organizations like the Machine Intelligence Research Institute and the Future of Humanity Institute focus on AI safety and studying potential existential threats.

What ACTUALLY is Effective Altruism?

It’s hard to know, but asking who is in the movement is an important starting point. Here are some results from within the movement, namely the EA outfit Rethink Priorities:

• The composition of the EA community remains similar to last year, in terms of age (82% 34 or younger), race (76% white) and gender (71% male).
• The median age when EAs reported getting involved in the community was 24.
• More than two thirds (69%) of our sample were non-students and <15% were undergraduates.
• Roughly equal proportions of non-student EAs report being in for-profit (earning to give), for-profit (not earning to give), non-profit (EA), non-profit (not EA), government, think tank/lobbying/advocacy careers.
• More respondents seem to be prioritizing career capital than immediate impact.

Thus, the movement is primarily young white men who are “prioritizing career capital,” meaning they are making a theoretical pledge to contribute in the future, after their financial ascendancy.

For those enticed to dive a bit deeper on EA, check out the book The Good It Promises, The Harm It Does, a scholarly blend of critique and analysis.

EA Loves Data

EA prides itself on a data-driven approach to philanthropy. It’s grounded in the belief that analytical and empirical analysis can lead to optimized charitable giving.

However, a core challenge emerges: the definition of “most good” remains ambiguous and open to interpretation. From my observation, the EA movement generally evaluates charitable actions based on:

• Addressing large-scale problems
• Enhancing the well-being of others
• Prioritizing initiatives where additional resources can make a significant difference

Unfortunately, this strategy may not account for the curse of dimensionality.

The Curse of Dimensionality

In layperson’s terms, the curse of dimensionality refers to the complications that arise when increasing the number of variables or dimensions in a given situation. To illustrate this concept, let’s do thought experiment.

Imagine yourself as an effective altruist aiming to maximize the welfare of individuals. In our hypothetical world, identity is measured by a single variable, distributed along a bell curve, which is also known as a normal distribution.

The individuals near the center are less marginalized, and it seems logical to direct resources towards this majority. If we target our resources at those within one standard deviation of the center, we reach approximately 68% of the population. Applause for efficiency, right?

But identity is not a single variable. Let’s now imagine identity as having two dimensions, still obeying the shape of a bell curve.

In this expanded scenario, targeting those within one standard deviation out from the center would reach only 39% of individuals.

Adding a third dimension of identity drops this figure to 20%. A fourth brings it down to 9%. A fifth, a mere 3%. With six dimensions, we’re only capturing 1% in that one standard deviation. This is a cool thing about math: dimension has a big impact on what shapes look like! As the dimensions of identity increase, the “fat middle” of the curve thins out drastically. Each added dimension underscores the significance of those dwelling in the margins — the long tail of the distribution.

This thought experiment is not a perfect reflection of reality; it’s a toy scenario. How many axes of identity are there? It depends how we define identity. Are those identities all distributed along a bell curve? Assuredly not. Nonetheless, the curse of dimensionality is ubiquitous. It is always going to be easiest to target folks who are less marginalized, because definitionally, they are a more homogeneous and easy-to-reach group. The sum of marginalized people might be greater, but they are marginalized in their own, unique intersectional ways. Supporting them requires hard, individualized, tailored work.

In short, EA’s goal of “doing the most good” sounds laudable but sidesteps the hard work of coping with intersectionality. EA gives up on helping some of the most marginalized — and likely most vulnerable — people because, well, they are hardest to help.

Conclusion

I certainly do not claim that all causes supported by Effective Altruism (EA) are unworthy of attention or funding. The global work done by the EA community on issues like malaria is commendable and has undoubtedly made a positive impact.

However, it’s crucial to examine the movement’s limitations. In the realm of social justice, particularly within the United States, the data-driven and efficiency-centric approach of EA can fall short. The complex, nuanced, and multidimensional nature of social justice issues doesn’t always align with EA’s methodology, and can be too easily deemed inefficient.

If you are part of the EA movement, my argument isn’t going to convince you of anything. You might come back at me with more diagrams of bell curves and some integrals and an Econ 101 discussion of utility functions and explanations of how I don’t understand your philosophy. And to you, I would say: why is EA so white and so male? And: do you have social justice outcomes you can show me?

Regardless, with prominent EA advocate SBF facing legal scrutiny, it’s time for renewed reflection on the composition and focus of the EA movement. And it’s essential to remember that true altruism includes helping those who are most overlooked.

Your nieghbor,

Chad