Research Grants

 Academic Paper Grant


 AI Risks Philanthropy: How Many Lives Can We Save per Dollar?


Research summary:

It appears that research into AI risk reduction strategies offers almost unprecedentedly high returns in expected lives saved per dollar. The proposed paper would spell out the case for this claim in enough detail so that outsiders or newcomers can check the evidence and arguments. The result would be submitted to a peer-reviewed journal.


Planned contents include:

  • A review of the notion of “optimal philanthropy”, of the literature to date on the returns (in lives saved, and similar measures) from various philanthropic endeavors, and of the large gaps in existing estimates of the returns from different endeavors.
  • A review of the literature on the value of reducing existential risks, and its limited connections to that on optimal philanthropy
  • Discussion of the virtues of back-of-the-envelope estimations procedures, even in cases with large unknowns.
  • Several different estimates of the returns from research into AI risks, reckoned by different methods. Discussion of structural uncertainty and of the merits of averaging the results of different models. A resulting estimate.

Prior related work:

Anna Salamon's 2009 Singularity Summit Talk "How much it matters to know what matters: A back of the envelope calculation" gives a quick overview of this type of estimation.


A handful of authors have done related work addressing human extinction risks generally (Matheny, 2007, Posner, 2004), on the estimated risk from nuclear catastrophe, and the huge gaps in that analysis (Hellman, 2008), and on the costs-benefit calculation for climate change prevention efforts given uncertainty about unlikely extinction-level impacts (Weitzman, 2007).


Venues for publication:
  • We will submit a paper to "Risk Analysis" or to another journal that has an interest in risk estimation and cost-benefit analysis.


Target dates for:

Extended abstract (Posting an extended abstract on SIAI website, and circulating to a few related academics for comment):2 weeks after starting date.

 

Working paper (Posting a working paper on the SIAI website; circulating to related academics): 10 weeks after starting date.

 

Journal submission: 18 weeks after starting date.


Follow-up steps (Brainstorming, and drafting proposals for, any follow-up publications.  Are there related research papers that should be considered?): 19 weeks after starting date.

Total budget:  $7,200

How research and writing costs are estimated:

  • Person-months for research and writing: 3 (our standard estimate[1] of 3 person-months per journal paper).
  • Dollars required to support one skilled full time researcher-month[2]: $2,400

[1] Our base estimate is 1.25 person-months per conference paper, and 3 per journal article, for an experienced full-time researcher. This estimate takes the planning fallacy, and the importance of an outside view in avoiding that fallacy, into account. While typical rates of article production by professors are extremely low, the distribution is strongly skewed towards research-oriented universities and departments, and informal surveys of researchers working on existential risks give data consistent with this estimate for full-time work required per paper. Visiting Fellows vary in their experience levels, so that mean productivity is expected to be lower, but a team mix can be selected to account for this.

[2] This billing rate reflects an estimate of financial outlays for SIAI to create the equivalent of one full-time skilled researcher-month, including stipend or hosting expenses, workspace, and administrative or management time, and other supporting expenses. Actual person-months may be greater or lower depending on the labor mix for a particular project, with shortfalls made up from general funds. This rate is not reflective of the money researchers could earn in the competitive labor market. Think of this as a matched donation. You donate the living expenses; our researchers donate the surplus value of their labor.

How this paper will help with existential risk:


Research benefits: (What ideas will the paper explore?  How will that knowledge help reduce existential risk?)
  • We may notice something major we’ve been missing in terms of the expected returns in existential risk reduction per dollar from AI risks research.
  • We may improve our understanding of what kinds of endeavors reduce risk.
  • We are likely to improve our estimation procedures for this estimate and for future, similar strategic estimates.

Influence benefits:(What target audience will the paper impact, how?  How will that impact help with existential risk?)
  • The paper will make it easier for unfamiliar academics and others to see the whole case for AI risk reduction efforts in peer-reviewed form.
  • It will help identify the areas of uncertainty in the argument where information value is high, encouraging additional research on those questions.
 

Human capital benefits, or network benefits (Will writing this paper help new Visiting Fellows become familiar with key research domains?  Will it help create relationships with outside co-authors?  Will it give folks interested in existential risk entry into new communities where valuable contacts may be found?)

  • Optimal philanthropy, cost-benefit analysis, and risk analysis communities are natural allies of AI risk reduction. So far we've had relatively little contact with these academics. This paper is an opportunity to cite the papers of such researchers, dialog with them, and lay foundations for possible future collaboration.
  • Co-writing this paper can help a Visiting Fellow critically examine the core case for AI risk reduction.



Credit card transactions are securely processed through PayPal.

PayPal – Grant Donation

(200 characters max)