Better Games or Bednets?

By Davidmanheim @ 2025-02-26T00:51 (+30)

I want to make the argument that, first, there is a plausible moral case that digital systems addiction and the consequent suboptimality is an enormous and ongoing tragedy for humanity, second, that the tragedy is important, neglected, and tractable, and finally, that this is all crazy, but plausible under many key assumptions underlying other effective altruist causes, and about what the implications of rejecting the claim might be. 

Draft Amnesty: This is a piece I mostly drafted three years ago, and never got around to finishing or posting - so I had claude help with parts of the second half, in order to finish it quickly. It is intended  to provocatively spark discussion, not to strongly argue for the conclusion that making video games is better than buying bednets - though I think it's not a crazy claim. (I debated posting this for April Fools, but draft amnesty week seems more appropriate, given that it's not actually joking.)

Digital systems addiction and suboptimality

Much of the developed world spends a significant amount of leisure time interacting with digital systems which are designed to be addictive and profitable to the operators, regardless of the impact on the users. In this discussion, the damaging digital systems include social media, video games, and mobile gambling.

For the case that this is a moral tragedy, it is not enough to say that these are a waste of time - if people enjoy wasting time in a certain way, their preferences should matter. Modern video games do not (necessarily) make people particularly socially isolated, and they can include healthy exercise, with examples including Wii Sports and Beat Saber. Instead, we will make the case that the entertainment they provide is sub-optimal from the perspective of users, and that the degree of suboptimality is significant.

Digital systems usage and damage.

In the United States, 75% of people play video games, for an average of around 15 hours per week. Some of the games that are played are horrible for mental health and user finances. I am very uncertain if roughly similar numbers apply around the world, but will guestimate that there are somewhere between 1-5 billion hours playing video games per day. (In comparison, There are approximately 3 billion monthly average users of Facebook, who spend an average of 30 minutes per day on the site. Followed by Youtube, Instagram, Tiktok, etc. I would estimate that there is very roughly the same amount of time spent on social media, but fixing that seems harder.)

These systems are designed for profit and to be addictive, implicitly if not explicitly. It seems overwhelmingly likely that replacing the most addictive video games with ones that are optimizing for a combination of fun and positive valence, using profitability as a constraint instead of a goal, would improve user experiences greatly.

Tractability and Cost-Benefit

Lots of people make video games. Indie developers do this all the time, with moderate levels of investment. Some portion are outrageously successful, with the positive externality of profits that could be spent on other beneficial causes.

Specifically, building better mobile games seems like a high risk / high reward intervention. Global mobile gaming revenue is around $100 billion annually, and gacha/loot box revenue is $30-50 billion annually. The cost of developing a new mid-tier video game is on the order of $10m. Claude's estimate is that for every $1 million invested in developing non-exploitative games, it could redirect $3-10 million from predatory systems, reach 100,000-1,000,000 players, create 50,000-500,000 hours of positive engagement, and has well-being improvements valued at $2-20 million (using standard well-being economic metrics.)

The direct benefits of this are clearly an order of magnitude or two less than bed-nets, but it would be directly profitable in ways that could fund other interventions, rather than only having longer term benefits.

Image / Perception

EA is already a game studio, and their reputation is only slightly worse than Effective Altruism. And isn't it about time that we start stealing the abbreviation back?

Conclusions

The comparison between video games and, for instance, anti-malarial bed nets is interesting; the raw cost-effectiveness numbers favor traditional interventions, but I claim that several considerations warrant deeper examination, namely, transferability, quality of life, and very high neglectedness.

First, the transferability of resources creates a more complex calculation than direct impact comparison suggests. If creating better digital systems is profitable, these profits could potentially fund traditional EA interventions at scale. A successful game studio generating $50 million in annual profit could theoretically fund millions of bed nets while simultaneously reducing digital harm. Second, we may be systematically undervaluing quality-of-life improvements in developed nations. If 2 billion people each waste just 30 minutes daily on suboptimal digital experiences, that represents over 365 billion hours annually of slightly diminished human flourishing. The cumulative scale of this welfare loss could be enormous, even if the per-person impact is small. Third, global health interventions receive billions in funding and attention, but positive-wellbeing games remains relatively unexplored. The field is wide open for innovation with potentially massive reach.

Many EAs accept that small improvements to animal welfare multiplied across billions of animals justify major interventions, but I think we would hesitate to apply similar logic to small improvements in human digital experiences multiplied across billions of relatively wealthy users. If we reject this intervention, are we being consistent in our moral reasoning, or are we privileging certain forms of suffering or certain populations? There is something morally unpalatable about focusing on "First World Problems" like bad video games - but are we embracing the claim that literal life-saving is fundamentally incomparable to mild experience enhancement, even if the costs and scale favors the later?