Few advancements have contributed to human progress as much as the Scientific Revolution, which paved the way for the Enlightenment and Industrial Revolution. The impact of science is long-term in nature, such that much of the benefits are attained by future generations whose world and worldview are shaped by the scientific output of the past. It is due to the immense power of science and the resulting pace of change in our world, however, that we must continually ask how science can evolve and mature to better serve our values, mitigate risks, and provide for our future.
Illustration adapted from Luke Mühlhauser1
Despite its immense track record, science is a largely fragmented system spread across nations and disciplines. It’s rife with inefficiencies (e.g. duplication of efforts, publication bias2, and replicability crisis3) while the distribution of resources in science poorly reflects common ethical goals4. Globalization, open science, and effective altruism have presented unique opportunities to optimize the pursuit of knowledge for the global good.
In recent decades, the adoption of cost-utility analysis5 in healthcare and the triage of philanthropy based on cost-effectiveness6 in effective altruism, have exemplified the effectiveness of expected utility in guiding decision-making under moral uncertainty7. Their effectiveness lies in the idea that some courses of action are hundreds or thousands of times as cost-effective as others by metrics of utility (eg. cost-per-life-saved). In healthcare, patient lifespan and quality of life have been enhanced even in highly resource-constrained healthcare systems by selecting medical interventions with the greatest cost-utility. In philanthropy, more than a billion dollars8 have been channeled to the world’s most cost-effective charities in global health and poverty, estimated to have averted hundreds of thousands of deaths.
However, I believe that there is even more potential impact to be gained from applying such concepts in the scientific community than in the world of philanthropy. Firstly, global gross domestic expenditure on research and development (GERD)10 is estimated at around $2 trillion annually (~2% of global GDP) and is, therefore, a much larger resource to leverage for the public good in comparison with philanthropic expenditures (~$150 billion annually)11. There is also an unparalleled talent pool in the scientific community, with much-coveted technical expertise that would take a huge shift in recruitment and several years for most philanthropic organizations to fill. Most importantly, as the impact of science is long-term in nature, it is best positioned to serve society’s long-term needs and ethical goals, benefitting generations far into the future.
Even within the applied sciences, the sciences closest to direct impact, large disparities have been observed in the distribution of scientific resources against value standards. For example, a report from the National Institutes of Health (NIH)4 showed that the distribution of research funding across diseases poorly reflects global mortality and disease burden, with pneumonia and hepatitis C receiving similar funding despite nearly a thousand-fold difference in disease burden. In a rational world, as per the concept of maximizing expected utility in effective altruism, the distribution of science funding would reflect societal needs to optimize for marginal utility.
Both basic science and applied science could be honed from the perspective of utility. Basic science acts as the knowledge-creation engine, with the downstream (often very distant) expectation that much of this new knowledge will serve as a foundation for generating value. Applied science acts as the filter, where knowledge is distilled to generate utility through verifiable solutions to problems. In reality, basic and applied science are much more connected with this, with most research falling into a middle-ground of use-inspired basic research (coined Pasteur’s Quadrant by Donald Stokes13). Thus the scientific process can be compared to directed evolution, whereby basic science introduces novelty and applied science is the selective pressure for utility.
For example, the 1000 genomes project14 provided an unprecedented catalog of genetic variation (basic) to mine for clinically relevant genetic insights into diseases (applied). Marginal utility is likely to be highest when knowledge gains are large, in an understudied field, and relating to questions of high ethical concern.
Experimenting with how scientists choose what problems to work on and how funding is assigned to better address global needs and achieve long-term ethical goals has enormous potential. There has already been some work in this direction by leveraging network science to study the mechanisms underlying scientific progress15, and early efforts to predict research impact in fields such as the behavioral sciences16 based on estimations of interventions’ impact on wellbeing. This presents the opportunity to re-evaluate science funding and coordination mechanisms to better support scientists in pursuing impact-driven science.
As science is global and multidisciplinary in nature, diverse input is required to achieve this goal. Building on globalization and developments in open science (which make the scientific output a globally accessible good), a coordinated effort among the scientific community is required to refine impact-driven science and empower scientists to pursue their most impactful science. PROMETHEUS, Promotion of Ethics in Science, is a grassroots movement towards this aim.
The goals of PROMETHEUS include:
- Build and coordinate an international community of scientists, science policymakers, and moral philosophers in pursuit of impact-driven science
- Identify quantifiable standards to guide research directions toward impact for both basic and applied sciences
- Identify imbalances in the distribution of resources between research topics against these standards
- Explore mechanisms of science funding and coordination that could better enable impact-driven science
- Empower scientists to pursue their most impactful science
The world of impact-driven science is vast, with many actors and stakeholders. This complexity should not hold us back but rather encourage us to be cautious, inquisitive, and incremental in pursuit of these ambitious goals. You can play a part in this mission. If you are interested in learning more about the movement or would like to get involved, check out our website or get in touch! We hope that as a community we can set the stage for a positive, impactful, and enduring chapter in the progress of science.
References:
- Mühlhauser, L. How big a deal was the Industrial Revolution? https://lukemuehlhauser.com/industrial-revolution/. ↩︎
- Thornton, A. & Lee, P. Publication bias in meta-analysis: its causes and consequences. J. Clin. Epidemiol. 53, 207–216 (2000). ↩︎
- Romero, F. Philosophy of science and the replicability crisis. Philos. Compass 14, (2019). ↩︎
- Report on NIH Funding vs. Global Burden of Disease. https://report.nih.gov/report-nih-funding-vs-global-burden-disease. ↩︎
- Robinson, R. Cost-utility analysis. BMJ 307, 859–862 (1993). ↩︎
- GiveWell. Cost-Effectiveness. GiveWell (2017). ↩︎
- MacAskill, W., Bykvist, K. & Ord, T. Moral Uncertainty. (Oxford University Press, London, England, 2020). ↩︎
- GiveWell’s Impact. GiveWell https://www.givewell.org/about/impact. ↩︎
- Caviola, L. et al. Donors vastly underestimate differences in charities’ effectiveness. Judgm. Decis. Mak. 15, 509–516 (2020). ↩︎
- OECD. Main Science and Technology Indicators. https://stats.oecd.org/Index.aspx?DataSetCode=MSTI_PUB. ↩︎
- Johnson, P. D. Global philanthropy report : Perspectives on the global foundation sector. (2018). ↩︎
- Global private and public R&D funding. Scienceogram UK https://scienceogram.org/blog/2013/05/science-technology-business-government-g20/ (2013). ↩︎
- Stokes, D. E. Pasteur’s Quadrant: Basic Science and Technological Innovation. (Brookings Institution Press, 2011). ↩︎
- 1000 Genomes Project Consortium et al. A global reference for human genetic variation. Nature 526, 68–74 (2015). ↩︎
- Fortunato, S. et al. Science of science. Science 359, (2018). ↩︎
- Lieder, F., Gainsburg, I., SamNolan & Corwin-Renner, E. Predicting the cost-effectiveness of future R&D projects and academic research. ↩︎