This page lists definitions of terms and ideas we sometimes use in our work. We wrote it to help authors at Open Phil provide references to their readers.
Artificial general intelligence
Computer program(s) that can perform virtually any cognitive task as well as any human. For more, see here.
Back-of-the-envelope calculation (BOTEC)
Often used to roughly estimate the expected cost-effectiveness of a potential grant, but can also be used to make other estimates. For more on how we use BOTECs, see here (under “Evaluating a grant’s cost-effectiveness”).
Bayesian mindset
An approach to quantifying uncertainty and using the resulting quantification to guide decision-making. For more on the Bayesian mindset, see here.
Calibration
The extent to which your statements (including predictions) adhere to the confidence level you assign to them. For example: if the predictions you make with 30% confidence are true about 30% of the time, you are well-calibrated. See here for further discussion.
Cause prioritization
The practice of researching which philanthropic causes offer the most benefit for the marginal investment. Said another way: meta-level philanthropic research that compares broadly different philanthropic causes. For more on cause prioritization, see here.
Cause selection criteria
Open Philanthropy evaluates causes according to three criteria: importance, neglectedness, and tractability. See more on our I/N/T framework here.
Importance:
How many individuals does this issue affect, and how deeply? How much good could a major breakthrough or victory accomplish? “Good accomplished” might include economic value created, income gains, improvements in health, reductions in suffering, increased odds of human flourishing, reduced odds of global catastrophes, and more.
Neglectedness:
All else being equal, we prefer causes that receive less attention from others, particularly other major philanthropists. Are there important aspects of a cause, or opportunities to make a difference, that receive little support relative to their importance? When investigating a cause, we tend to consider multiple different kinds of activities that might make a difference. Even within a well-funded cause, some activities or sub-areas might be relatively neglected.
Tractability:
We prefer causes for which there are clear ways a funder could contribute to significant progress. It can be difficult to anticipate what opportunities will arise and how long they will last, but there are some issues where we see relatively broad and robust dynamics that make progress particularly likely — or unlikely.
Corporate campaigns
An effort to change corporate practices with a carrot-and-stick approach. Common methods include supporting aligned stakeholders or offering technical assistance (carrot) and launching social media campaigns against corporations that don’t engage (stick). For more on corporate campaign use (specifically in animal welfare efforts), see here.
Counterfactual
Relating to or expressing what has not happened or is not the case. (I.e., “without the support of Philanthropy X, we estimate that it would have counterfactually taken Y more years for The Pill to become widely available in the US.” For more, see here.
Disease burden
The impact of a health problem as measured by financial cost, mortality, morbidity, or other indicators. Often measured via disability-adjusted life years (DALYs) and years of life lost (YLLs). For more on disease burden, see here.
Forecasting
The practice of making probabilistic predictions about an outcome. (For example, “I’m 70% confident the grantee will achieve milestone X within 1 year.”) Read more about our internal predictions here, and about our dedicated Forecasting cause area here.
Global Catastrophic Risks (GCR)
Used to refer to the broad range of work within OP’s GCR portfolio. In GCR work, we often ask questions like: “How much does this grant reduce the chance of a catastrophic event that endangers billions of people?
Cause areas within this portfolio include: Biosecurity and Pandemic Preparedness, Forecasting, Global Catastrophic Risks Capacity Building, and Potential Risks from Advanced Artificial Intelligence.
A similar term, “existential risk”, also appears in some of our work. Existential risks are a subset of global catastrophic risks — those which could “threaten the destruction of humanity’s long-term potential” by causing human extinction or another disaster from which humanity couldn’t recover (like an irreversible collapse of civilization or permanent global takeover by a totalitarian regime).
We formerly referred to this portfolio as “Longtermism”. We changed the name to reflect our view that the risks we focus on aren’t only relevant in the long term; they could threaten the lives of many people in the near future. (The current name is also a better parallel for “Global Health and Wellbeing”.)
Global Health and Wellbeing (GHW)
Used to refer to the broad range of work within OP’s GHW portfolio. In GHW work, we often ask questions like: “How much does this grant increase health (denominated in, e.g., life-years) and/or wellbeing, worldwide?”
See full list of cause areas within this portfolio here.
Grant tiers
A system for ranking our GCR grants when we recommend them; tier 1 is the highest rank, while tier 5 is the lowest rank we’re open to funding. See this post for more on the tier system.
(Our GCR portfolio used to be called “Longtermism”; that name is used in the above post.)
Hits-based giving
An approach to philanthropy that is guided by expected value. This means an unusual openness to supporting work with high upside and high chance of failure. For further discussion, see here.
LMICs (Low- and middle-income countries)
A country that meets the criteria for being “low” or “middle” income. Open Philanthropy generally follows the World Bank’s classification system.
Longtermism
Roughly speaking, an approach to doing good based on the premises that:
- Future people matter.
- There could be a lot of them.
- We can make their lives better or worse.
For more, see here.
Marginal impact
The impact of adding an additional dollar of funding. Due to the phenomenon of diminishing returns, the impact of each additional dollar will decrease after a certain point. For further discussion on marginal impact, see here.
Measurements of health outcomes (DALYs, YLLs)
DALYs (Disability-adjusted life years)
One DALY represents the loss of the equivalent of one year of full health. This metric is often used to assess the overall burden of a disease. OP values 1 DALY at $100,000. For more on DALYs, see here.
YLLs (Years of life lost)
A measure of premature mortality that considers both the frequency of deaths and the age at which it occurs. For more on YLLs, see here.
Modes of forecasting (inside, outside)
The two primary modes of forecasting developed by Kahneman and Lovallo (1993).
Inside view:
An inside view focuses on the specifics of the case at hand, relying on detailed information about the project. For more on the inside view, see here.
Outside view:
An outside view takes a broader perspective, looking at the outcomes of similar cases to estimate the likely outcome of the current project. For more on the outside view, see here.
Moral patient
A being that deserves moral concern. For more, read here.
Moral weight
Roughly, the amount of importance we give to the well-being of different kinds of moral patients. For more on moral weights (specifically as they relate to non-human animals), see here.
Normative uncertainty
Uncertainty about morality, particularly core aspects of morality that could dramatically change an expected value calculation. (I.e., do you value human life and chicken life equally? Do you value human life 10-100x more?) For more, read Holden Karnofsky’s essay on worldview diversification, which explicitly discusses this problem.
Open Philanthropy’s “bar” (or “The bar”)
Within our Global Health and Wellbeing (GHW) portfolio, Open Philanthropy uses a “cost-effectiveness bar” to compare different grants to one another and ensure that grants meet our standards for how impactful (in expectation) they are. We compare our back-of-the-envelope cost-effectiveness estimates for our grants to the bar and aim to make grants where our modeled cost-effectiveness is higher than our bar.
Our bar is dynamic and has changed over time. Most of our GHW programs, we use a bar that is equivalent to averting a disability-adjusted life-year for around $50 or increasing the income of four people income by 1% for a year for $1. Our Farm Animal Welfare program uses a separate bar.
For a more technical description of the bar, see here. For a briefer, more recent discussion that mentions our GHW bar as of 2024 (slightly over 2000x), see here.
OP value/OP dollars
How much Open Philanthropy values a benefit in USD. A DALY is worth $100,000 in OP value: if a charity can avert a DALY for $100, we’d think of it as having a 1,000x return, because the benefits would be $100,000 relative to the cost of $100.
A doubling of income is worth $50,000 in OP value regardless of the person’s income. We’re concerned with marginal impact (the impact of an additional dollar spent), and we treat a dollar of extra income as being worth 100x more to someone who has 100x less.
These numbers have changed throughout Open Phil’s history, and they may continue to do so. For a more detailed discussion of how we’ve arrived at these values, see here.
Randomized controlled trials (RCTs)
A form of impact evaluation in which the population receiving the program or policy intervention is chosen at random from the eligible population, and a control group is also chosen at random from the same eligible population. For more, see here.
Reasoning transparency
A set of writing practices that pushes the writer to communicate more transparently. Writers are encouraged to:
- Open with a linked summary of key takeaways.
- Indicate which considerations are most important for [their] key takeaways.
- Indicate how confident [they] are in major claims, and what support [they] have for them.
For further discussion, see here.
Social return on investment
The estimated social benefits per dollar of cost for a specific program or intervention. The social benefits include values not included in a private ROI, e.g., social, environmental, and economic outcomes. For more on SROI, see here.
Transformative AI
Potential future AI that precipitates a transition comparable to (or more significant than) the agricultural or industrial revolution. For further discussion, see here.
Updating
Broadly used to describe the practice of changing one’s stance on something in light of new information. (For example, “I estimated a 70% chance the grantee would achieve their goals, but in light of new information, I’ve updated downwards.”) Falls under the umbrella of Bayesian thinking.
Willingness to pay (WTP)
Open Philanthropy’s willingness to pay for an opportunity/grant. This is calculated by taking the OP value of an opportunity and dividing it by our bar.
Worldview diversification
The practice of putting significant resources behind each worldview that OP finds highly plausible. For further discussion, see here.