I am an attorney in a public-sector position not associated with EA, although I cannot provide legal advice to anyone. My involvement with EA so far has been mostly limited so far to writing checks to GiveWell and other effective charities in the Global Health space, as well as some independent reading. I have occasionally read the forum and was looking for ideas for year-end giving when the whole FTX business exploded . . .
Jason
Karma: 17,530
Most of this seems unrelated to my observation that evidence of weariness toward harshly criticizing people merely for their political beliefs does not provide much evidence of weariness for criticizing extremely powerful political people for their actions.
Yes, I know what due process is. In a strict sense, it applies only to governmental actions. In a looser sense, it applies to actions by private organizations that are considering taking adverse action against people associated with it—e.g., a private university which is considering disciplining a student, a homeowner’s association which is considering fining a unitowner. I am not aware of its extension to citizens (or groups of citizens) who wish to criticize a senior government official, or to a social movement that wishes to express its disapproval of a prominent person and distance themselves from said person. There would be far less speech if the so-called “court of public opinion” acted with the constraints of government action.
Even if due process did somehow apply, the classic case on the subject explains that “identification of the specific dictates of due process generally requires consideration of three distinct factors: first, the private interest that will be affected by the official action; second, the risk of an erroneous deprivation of such interest through the procedures used, and the probable value, if any, of additional or substitute procedural safeguards; and, finally, the Government’s interest, including the function involved and the fiscal and administrative burdens that the additional or substitute procedural requirement would entail. ” Mathews v. Eldridge, 424 U.S. 319, 335 (1976).
The private interest of Elon Musk in not being criticized by EAs is exceedingly modest. I think the expected value of additional safeguards is low—he has made numerous troubling, divorced-from-reality, or erratic public statements and I am unaware of any dispute about whether the statements are his. Finally, with the volume of problematic statements, and new ones coming out on a near-daily basis, the burden of adjudicating them all would be considerable for a relatively small social movement.
Relatedly, one of the reasons we can hold the government to the dictates of due process is that we concurrently give it the powers needed to carry out its functions while providing more protections. EA cannot subpoena Musk to judge him, or force him to sit for depositions on pain of imprisonment. Nor can it seek funds from the public fisc for investigations.
Indeed, the American constitutional tradition gives people like Musk less protection than you or I: he is a quintessential public figure for defamation purposes. Owing to his control of X, he has an ability to respond to criticism that is exceptional even in comparison to most public figures.
The isolated demand for “due process” rigor here is somewhat ironic given Musk’s habit of shooting from the hip with criticisms about others, including federal employees when he is a public official, employment reform is within the scope of his apparent job functions, and people are suffering concrete harm (loss of jobs and contracts) as a result of his official conduct on the topic. Further irony comes from the fact that he owns a major social media platform that is infamous for people offering uninformed hot takes. These factors do not affect how we should treat him, but the irony is palpable.
There may be prudential reasons for treating Musk with kid gloves, or for not talking about him. “Due process” ain’t it.
Assuming there’s effective political stuff to be done with respect to the USAID situation (which is uncertain to me), it’s plausible that any hint of EA involvement would be affirmatively counterproductive. Better to have more politically popular entities—and entities not predominately funded by a guy who gave megabucks to the current officeholder’s rivals—in the lead for this one. If, for instance, EAs wanted to funnel money to any such entities, I suspect it would be savvy to do so quietly rather than talking about it on-Forum. It’s possible that is playing a role in the lack of discussion here, although I too suspect this would have gotten more attention ~18 months ago.
<<very fed up with the practice of denouncing people for their political beliefs.>>
Among other things, there’s a major difference between merely holding political beliefs and one’s actions as apparently one of the most currently influential people in the most powerful organization that has ever existed in history (i.e., the US government).
We actually did not include most of our analysis or evidence in the review we published, since brevity is a top priority for us when we write reviews. The published review is only a small fraction of the problems we found.
I’d suggest publishing an appendix listing more problems you believe you identified, as well as more evidence. Brevity is a virtue, but I suspect much of your potential impact lies with moving grantmaker money (and money that flows through charity recommenders) from charities that allegedly inflate their outcomes to those that don’t. Looking at Sinergia’s 2023 financials, over half came from Open Phil, and significant chunks came from other foundations. Less than 1% was from direct individual donations, although there were likely some passthrough-like donations recommended by individuals but attributed to organizations.
Your review of Sinergia is ~1000 words and takes about four minutes to read. That may be an ideal tradeoff between brevity and thoroughness for a potential three-to-four figure donor, but I think the balance is significantly different for a professional grantmaker considering a six-to-seven figure investment.
It’s a hypothetical about feeding homeless people, so it doesn’t track any of your prior decisions not to give advance notice to the target organization.
In the hypothetical, you could assert that advance notice wasn’t warranted because your external evidence was overwhelming enough that providing advance notice and opportunity for comment would have been a waste of time. This would be consistent with my interpretation of the community norm, which is that critics should provide advance notice unless they can (and do) articulate a good cause on a situation-specific basis to dispense with doing so. But if you did provide this justification, and the organization was later able to provide a response that materially changed the conclusions to be drawn, then your organization would suffer a loss in credibility as a result.
I think your second strand of evidence is doing almost all the work here. There’s a big difference between evidence sufficient to establish a prima facie case, such that a reasonable mind would accept the conclusion if no additional evidence were offered and evidence that so conclusively establishes the truth that it’s a waste of time to ask the charity for a response.
As with other considerations, this is a fact-bound consideration. If you’re relying on it to justify not giving a specific organization notice and an opportunity to prepare a response for concurrent publication, you can always say that—and thereby accept the loss in your own credibility if readers later judge that the charity had a reasonable response that it could have provided if asked.
I would include situations in which the charity could establish the substantial truth of the assertion. E.g., if the charity could show that it made a math error and actually fed 9,823 individuals, that is information a reasonable reader would find highly relevant to evaluating the assertion of falsity. Not giving the charity a chance to concurrently explain that does a disservice to both the charity and the reader.
I think a reasonable reader would view these statements as assertions grounded in a cost-effectiveness estimate, rather than as some sort of guarantee that Singeria could point to 1,770 piglets that were saved as a result of my $1 donation. The reader knows that the only plausible way to make this statement is to rely on cost-effectiveness estimates, so I don’t think there’s a meaningful risk that the reader is misled here.
It appears we agree that Sinergia is making false claims about helping animals.
I can’t speak for Aidan, but the word false has certain connotations. Although it doesn’t explicitly denote bad faith, it does carry a hint of that aroma in most contexts. I think that’s particularly true when it is applied to estimates (cost cost-effectiveness estimates), opinions, and so on.
I think that phrasing should be used with caution. Generally, it would be better to characterize a cost-effectiveness estimate as overly optimistic, inaccurate, flawed, or so on to avoid giving the connotation that comes with false. False could be appropriate if the claims were outside the realm of what a charity could honestly—but mistakenly—believe. But I don’t think that’s what Aidan was saying (that reading would imply he thought “most charities[]” were guilty of promoting such claims rather than merely ones that were overly optimistic).
This rationale only works if you very carefully calibrate your claims. “Charity X has not provided sufficient, publicly available evidence to substantiate its claim to have provided 10K meals to homeless individuals” is a much different claim than “Charity X didn’t feed 10K meals to homeless individuals.” The former is about transparency; the latter is about whether the claimed results are true or not. An assertion that lack of transparency is not acceptable does not warrant an inference linking a lack of transparency to the falsity of the claim—at least not without giving the charity a chance to correct its alleged transparency error ahead of time.
It’s not difficult to try. I believe it would be difficult to succeed (or even not get easily caught) because the evidence exists in too many third-party hands outside the organization’s custody and control. Thus, absent special circumstances, it is very unlikely an organization would even try; the difficulty of success would deter them.
(I meant “risk” to refer more to the consequences of being caught—that it would pose a grave risk to the organization’s continued existence, and to the careers of those who attempted the tampering.)
Risk 1: Charities could alter, conceal, fabricate and/or destroy evidence to cover their tracks.
While this is possible in the abstract, I don’t think it is generally applicable enough to credit without a case-specific application to the facts at hand.
For example: Your most recent post asserts that Singeria has made “false claims” and that this is an “important issue because Sinergia receives millions of dollars in grants from EA organizations.” You note specifically that it has been a ACE top charity since 2018 and recently received a $3.3MM grant from Open Phil. I’ll allude to that at points in the discussion as an example of what types of reports you might be producing, but emphasize that the discussion is entirely hypothetical.
One problem is that your argument for importance also underscores how difficult and risky for an organization to “alter, fabricate, and/or destroy evidence” in the way you describe. If the charity challenges your work, presumably you’d produce the screenshots. At that point, the charity would have to double down and accuse you of fabricating the screenshots; there’s no plausible way to claim “misunderstanding” at that point in the game.
That would be a serious charge, not least because I would expect (and generally hope) that grantmakers would impose an organizational death penalty on an organization that fraudulently claimed that an evaluator’s screenshots were forged. Stuff like misattributed credit and poor calculations can be written off as non-malicious error, but deceit about something as concrete as what the organization’s website said would be very hard to spin as anything other than career/organization-ending lack of trustworthiness.
In addition to general Internet archives, there are a number of places not under the organization’s control where corroborating evidence could potentially be found. And I think it would be very hard for an organization to reduce the risk of information leakage to a level it deemed an acceptable risk:
Presumably, the alleged false claims would appear in fundraising and fundraising-adjacent materials in the possession of third parties, such as Open Phil, ACE, other grantmakers, recipients of fundraising e-mails, etc.
It’s likely that references to the false claims would also exist in some form in individual memories and in written discussions under the control of disinterested third parties (e.g., discussions on the Forum, which I believe are regularly scraped by various people). Although this might not be conclusive, it would leave the organization under a cloud of suspicion probably worse than admitting the truth of the original allegations!
There’s also the angle that falsely accusing someone of forging screenshots would constitute libel—and even in the US, the organization isn’t likely to have any good defenses other than non-falsity of the accusation. And litigation would bring even more tools to help uncover the truth:
If a preservation demand were filed quickly enough, third-party service providers (e.g., a webhost, an Internet archive that received takedown requests) may have evidence establishing the attempted cover-up. Some of these might even voluntarily release information without a lawsuit even being filed.
The organization might have to turn over its own records in discovery that would establish the coverup. Of course, the organization could try to hide documents, but that comes with risks!
Of course, there might not be litigation . . . but I wouldn’t underestimate the risk to the organization. If there were enough suspicion, I think people would fund it if you were able to conditionally commit some funds to it upfront and/or if there were some third-party recollections of the disputed statements. The litigation would be fairly simple and thus cost-effective as a way to determine whether the organization should receive funds.
Given the difficulties and risk of accusing an evaluator of committing forgery, I’d generally be inclined to give this issue meaningful weight where there was a reasonable basis to believe that:
* Distribution of the information to third parties was limited enough that the organization could predict that evidence tampering had a high degree of success.
* The report was so damning that the organization had little to lose from claiming forgery (cf. the Hail Mary pass situation in American football).
* Organization leadership was irrational enough to claim forgery even when doing so was strongly against the organization’s interests and their personal interests.
To clarify the structural reasons to which I alluded—it’s already problematic for the community that OP et al. control so much of the available funding. Having mostly independent funding streams available is important, so having OP et al. putting a thumb on the scale to influence allocation of those independent streams poses significant costs in my book.
A necessary prerequisite for your proposal is that OP et al. move toward the cause prio you advocate. But such a movement would presumably cause OP’s own funding decisions to shift in a significant manner. OP et al. changing allocations of its own funds doesn’t pose the same problems to funding independence, and doesn’t require confronting the likelihood that EGI fundraising from outside sources is less cost-effective on a dollar basis in AW than in GHD.
It’s not clear to me whether the majority view that AW is currently underfunded would survive a significant reallocation in OP et al.’s allocation of its own funds. Even if it did, there would still be significant support in the community for the view that the marginal independent-funding dollar should go to GHD or GCR work (because there already is), and we’d still be facing the question of whether it was appropriate for OP et al. to be exercising pressure on this question.
While I think reasons (1) and (2) are relevant to the degree of importance/value in advance notification is necessary, I just don’t see any legitimate reason to rush to press here. If Singeria had just dropped new claims (especially during/just before the end of year fundraising period), or was in the public spotlight for some related reasons, then there would be a stronger argument for appreciable costs to delaying publication by a week. I don’t see any justifications like that offered. I think you need some meaningful alleged harm from delay or advance notification before mitigating factors like (1) and (2) could come into play.
Reason (3) is unexplained—why do you think hearing from the charity “can introduce biases”? Moreover, even if you believe that, I don’t see why you couldn’t at least send an advance copy to Singeria a week in advance with a pre-commitment that you were not going to change the text absent proof of a clear factual error. That would allow Singeria to prepare a response for submission concurrently with your critique, and for the reader to see both sides of the dispute at once. Choosing not to do so means that a decent number of readers who see your charges will not see Singeria’s response, a state of affairs that does not further truth discovery.
As for myself, I am increasingly inclined not to read substantive critiques of this sort absent either good cause for not providing an advance copy or the passage of enough time for the organization to respond. If my Forum viewing habits mean I don’t see the response (and don’t remember to reopen the post after a week or so), then that’s OK. I think the downside of missing some critiques is likely outweighed by avoiding the cognitive biases that can come with delayed presentation of the other side of the story and/or the risk of missing the other side when it comes out.
I worry EGIs’ cause prioritisation is ultimately restricted by Dustin Moskovitz and Cari Tuna. These are the major funders of Open Philanthropy (OP), which in turn is the major funder of EGIs, and therefore sets their incentives. I encourage people at OP to be deliberate about differences in cost-effectiveness between cause areas instead of incentivising EGIs to simply increase donations to top interventions regardless of their areas.
While I generally share concerns about OP et al. having too much influence over the ecosystem, that doesn’t sound like what you’re describing here. Rather, it seems that you think OP et al. are implicitly deferring to judgments made of the end users of EGIs? It seems to me that would be the main practical effect of using donations to effective causes as a metric without scoring differently by cause area. And I think there are some pretty powerful structural reasons we would want OP et al. to broadly defer to EGI customers rather than attempt to use its influence over EGIs to shape the ecosystem in furtherance of any particular non-consensus cause prio.
More substantively: What is your model of potential EGI end users? My assumption is that most people are coming in with a cause area they are interested in, and need to feel that the EGI aligns fairly closely with their values in order to trust it. I am guessing the number of prospects who wish to offload their cause prio and moral judgment to an EGI is much lower. To the extent that an EGI was heavily promoting AW opportunities over others, I submit that it needs to be upfront about being an AW EGI. Anything else is going to come off as a bait-and-switch to most potential end users. The downside is that the catchment area (as it were) for AW EGIs may be much smaller than for multi-cause EGIs that take all comers and don’t clearly prefer some of their cause areas over others.
I instead expect roughly the results of adding some pretty smart and hardworking people.
The usefulness of smart people is highly dependent on the willingness of the powers-that-be to listen to them. I don’t think lack of raw intelligence had much of anything to do with the recent US electoral results. The initial candidate at the top of the ticket was not fit for a second term, and was forced out too late for a viable replacement to emerge. Instead, we got someone who had never polled well. I also don’t think intelligence was the limiting factor in the Democrats’ refusal to move toward the center on issues that were costing them votes in the swing states. Intellectually understanding that it is necessary to throw some of your most loyal supporters under the bus is one thing; committing to do it is something else; and actually getting it done is harder still. One could think of intelligence as a rate-limiting catalyst up to a certain point, but dumping even more catalyst in after that point doesn’t speed the reaction much.
I think @titotal’s critique largely holds if one models EAs as a group as exceptional in intelligence but roughly at population baseline for more critical and/or rate-limiting elements for political success (e.g., charisma, people savvy). I don’t think that would be an attack—most people are in fact broadly average, and average people would be expected to fail against Altman, etc. And if intelligence were mostly neutralized by the powers-that-be not listening to it, having a few hundred FTEs (i.e., ~10% of all EA FTEs?) with a roughly normal distribution of key attributes is relatively unlikely to be impactful.
Finally, I think this is a place where EA’s tendencies toward being a monoculture hurts—for example, I think a movement that is very disproportionately educationally-privileged, white, STEM focused, and socially liberal will have a hard time understanding why (e.g.) so many Latino voters [most of whom share few of those characteristics] were going for Trump this cycle and how to stop that.
“AIDS related deaths in the next five years will increase by 6.3 million” if funding is not restored, UNAIDS executive director Winnie Byanyima said.
This is a quote from a BBC news article, mainly about US political and legal developments. We don’t know what the actual statement from the ED said, but I don’t think there’s enough here to infer fault on her part.
For all we know, the original quote could have been something like predicting that deaths will increase by 6.3 million if we can’t get this work funded—which sounds like a reasonable position to take. Space considerations being what they are, I could easily see a somewhat more nuanced quote being turned into something that sounded unaware of counterfactual considerations.
There’s also an inherent limit to how much fidelity can be communicated through a one-sentence channel to a general audience. We can communicate somewhat more in a single sentence here on the Forum, but the ability to make assumptions about what the reader knows helps. For example, in the specific context here, I’d be concerned that many generalist readers would implicitly adjust for other funders picking up some of the slack, which could lead to double-counting of those effects. And in a world that often doesn’t think counterfactually, other readers’ points of comparison will be with counterfactually-unadjusted numbers. Finally, a fair assessment of counterfactual impact would require the reader to understand DALYs or something similar, because at least a fair portion of the mitigation is likely to come from pulling public-health resources away from conditions that do not kill so much as they disable.
So while I would disagree with a statement from UNAIDS that actually said if the U.S. doesn’t fund PEPFAR, 6.3MM more will die, I think there would be significant drawbacks and/or limitations to other ways of quantifying the problem in this context, and think using the 6.3MM number in a public statement could be appropriate if the actual statement were worded appropriately.
But I don’t think people blindly defer to evaluators in other life domains either—or at least they shouldn’t for major decisions. For instance, there are fraudulent university accreditation agencies, non-fraudulent ones with rather low standards, ones with standards that are pretty orthogonal to whether you’ll get a good education, and so on.
I suggest that people more commonly rely on a social web of trust—one strand might be: College X looks good in the US News rankings, and I trust the US News rankings because the school guidance counselor thought it reliable, and the guidance counselor’s reputation in the broader school community is good. In reality, there are probably a couple of strands coming out from US News (e.g., my friends were talking about it) and from School X (e.g., I read about some successful alumni). So there’s a broader web to justify the trust placed in the US News evaluation, buttressed by sources in which the decisionmaker already had some confidence. Of course, the guidance counselor could be incompetent, my friends probably are ill-informed, and most schools have at least a few successful alumni. But people don’t have the time or energy to validate everything!
My guess is that for many people, GiveWell doesn’t have the outgoing linkages that US News does in my example. And it has some anti-linkages—e.g., one might be inclined to defer to Stanford professors, and of course one had some harsh things (unjustified in my opinion) to say about GiveWell. It comes up in the AI overview when I google’d “criticisms of Givewell,” so fairly low-hanging fruit that would likely come up on modest due diligence.
I’d also note that independent cannot be assumed and must be either taken on trust (probably through a web of trust) or sufficiently proven (which requires a fair amount of drilling).
My guess is that GiveWell is simply not enmeshed in John’s web of trust the way it is in yours or mine. Making and sustaining a widely trusted brand is hard, so that’s not surprising.
<I’m a bit disappointed, if not surprised, with the community response here.>
I can’t speak for other voters, but I downvoted due to my judgment that there were multiple critical assumption that were both unsupported / very thinly supported and pretty dubious—not because any sacred cows were engaged. While I don’t think main post authors are obliged to be exhaustive, the following are examples of significant misses in my book:
“animals successfully bred to tolerate their conditions”—they are bred such that enough of them don’t die before we decide to kill then. It’s hard to see much relevance to whether they have been bred to not suffer.
The focus on humans “failing even to reproduce at replacement rates” suggests a focus on wealthy countries. Where’s the evidence that people in wealthy countries are not reproducing due to too much suffering, or discontent with their alleged domestication? The post doesn’t engage with people in those countries being happier than the median; the effects of education, income, and changing social mores on fertility; and other well-known factors.
More fundamentally, the assertion that there are “similar dynamics” at play between humans and farmed animals is particularly unsupported in my view.
It’s important to not reflexively defend sacred cows or downvote those who criticize them . . . but one can believe that while also believing this post seriously missed the mark and warrants downvotes.
Yes, that’s correct.
And possibly the proposed cuts to the US defense budget, which I mention because those proposed cuts may be more likely to outlive Trump than the NATO stuff. (Maybe I am projecting because, as a US taxpayer, I have long felt that the US spends too much on defense.)