What questions relevant to EA could be answered by surveying the public?

David_Moss

[Question]

What questions relevant to EA could be answered by surveying the public?

Comments 19

Sorted by

New & upvoted

weeatquince🔸

Any human-focused moral weights work!!

How much do members of the public care about:

Subjective wellbeing
Increases in income
Increases in happiness
Reductions in pain
Mental health
Education
Being alive

Public surveys would be crucial for developing better QALYs / DALYs / WELBYs / etc (see these posts).

Public surveys are also needed to make trade offs between health and things not captured by QALYs / DALYs (such as increased income or justice), to trade of between years of life and quality of life (especially for some population ethics view) and so on.

Surveys in developing countries would be particularly useful.

Sanjay

SoGive conducts research like this as part of its moral weights process.

Our first such study was last year and focused primarily on how much people value

Saving lives
Doubling consumption
Averting severe depression
Averting animal suffering

We invested less effort in, but also explored

Saving species from extinction
Comparing a life in the far future with a life today
Education

We were interested in comparing them against each other, and in a quantitative comparison (i.e. how much is one valued more than another)

We were motivated to conduct this research because our work is intended to serve a broad audience, and we wanted to incorporate the perspective of the population as a whole.

Because the questions we wanted to answer were quantitative, we posed quantitative questions in the survey. We found that several respondents gave answers which did not seem consistent, when consistency checks where included in the survey. This is a key challenge for this type of work, and suggests that a slightly more indirect approach to answering the questions may be more effective.

calebp

I think that JPAL has done some work on this but I can't find the papers after a quick google search.

I am a bit confused why this is actually useful? Is it mostly for optimising for preferences, if so I can see why this would be useful but if you don't strongly prioritise preferences then I don't see why this would help you create better metrics?

I think my question is 'for people who don't prioritise preferences (e.g. hedonistic utilitarians), do you still think that moral weight surveying is useful? '? E.g. maybe an intervention will be more widely adopted by a community if appeals to their preferences, so efficacy is increased from a non-preference perspective.

David_Moss

Fwiw, I think that both moral uncertainty and non-moral epistemic uncertainty (if you'll allow the distinction) both suggest we should assign some weight to what people say is valuable.

Moral uncertainty may suggest we should assign some weight to views other than hedonistic utilitarianism. This includes other moral views , not just people's preferences, and we can discern what moral views people endorse through surveys (as I mention here). So we should ask about what people value and/or think is morally right and good, not merely what they prefer.
In addition, some moral views assign value to things which can be determined through surveys, including preferences (which you mention), but potentially including things like respecting people's values, autonomy/self-determination, democratic will, and not traducing people's wishes or coercing them.
But, separately, even if we only value maximizing wellbeing, given uncertainty about what promotes this / measurement error in our measures of it (and perhaps also conceptual uncertainty about what it consists in, though this may collapse into moral uncertainty), I think it's plausible we should assign some weight to what people say they prefer in judging what is likely to promote their wellbeing. For example, if we observe that having children seems to lead to lower wellbeing, but that people report that they value and prefer having children, that seems like it should be assigned some weight.

Pablo

I think that both moral uncertainty and non-moral epistemic uncertainty (if you'll allow the distinction) both suggest we should assign some weight to what people say is valuable.

Only ~7% of all people who ever lived are currently alive. What's the justification for focusing on humans living in 2022? Is it just that figuring out the values of past generations is less tractable?

David_Moss

It seems plausible that we should assign weight to what past generations valued (though one would likely not use survey methodology to do this), as well as what future generations will value, insofar as that is knowable.

JoelMcGuire

I agree that moral uncertainty implies it's a good idea to know what people's moral views are.

Related to your last point:

given uncertainty about what promotes this / measurement error in our measures of it (and perhaps also conceptual uncertainty about what it consists in, though this may collapse into moral uncertainty), I think it's plausible we should assign some weight to what people say they prefer in judging what is likely to promote their wellbeing.

Many EAs want to maximize wellbeing, and many pursue that aim using evidence. Given that, I'd be curious to know how views differ between experts, EAs, and the public on what wellbeing is and how can we measure it. I wrote a very rough example of the type of questions I could imagine asking in this document.

JoelMcGuire

I would be interested to know the results of such a survey on these topics.

Similarly, if experimental philosophy hasn’t already answered these questions, then I’d like to know if the public has any coherent views on “what wellbeing is, and what’s the badness of death?” I haven’t found anything in the literature that I could use, but I’m not very familiar with the research in this space. I think David has mentioned there being some extant literature surveying views on the badness of death, but I was not able to find it.

I know that David, you said:

I mostly intend to rule out things like surveying effective altruists or elite policy-makers.

But if we survey the public moral views, I'd like to know how much they differ from EAs. At the very least for communications

weeatquince🔸

What is the public's views, visions and ideas of what a good future will look like?

The idea here is that a clear vision of a what a good future looks like has been a key part of successful long-term policy making to date (based on experiences in Wales and Portugal). The hope is a clear vision of what the public want helps make long-term decision making feel easier to democratic policy makers, it helps them to explain and justify a focus on the long-term and should ultimately helps policy-makers prioritise the long-term more.

NoemiD

I've been playing around with the idea of doing something in this direction - if any collaborators want to work on doing some survey or experimental work exploring this, feel free to reach out!

weeatquince🔸

I am doing work on this in the UK. Will PM you.

Edit: I do plan to do some of this. So if anyone else is interested in helping with such work on the UK do let me know.

brb243

The issues here can be that

respondents can omit morally further individuals in their considerations
- so a good future for all sentient beings would have to be specified or skillfully implied by questions leading to this one
replies can only include what respondents can imagine, can be influenced by popular media portrayals, and may focus on solving personal problems
- for example, a person in a negative relationship who sees commercials that portray a product providing positive emotions may reply that people [like them] would have products to feel good - this does not identify the fundamental problems or include others' needs
  - thus, the ability to empathize with different alternatives, even those not commonly portrayed, should be offered and inclusivity in mental image invited
advanced economies' public may seek to secure a good future for itself, even risking the institutionalization of norms negative or suboptimal for others
- thus, this survey should be run in different parts of different economies
  - but then, especially in emerging economies, ways to mitigate experimenter bias should be found

So, I would pay attention to the questions specifics to make appropriate thought-through recommendations to policymakers.

The alternative is to have a broad solution, such as continuously improving the welfare of all sentience, and then ask the public to endorse it by a sequence of type/wording of questions or solicit answers that can accommodate the inclusion of consideration/can be interpreted as such (for example, 'do you want future good for all?' - yes!) - this should have much lower critical thinking requirements.

weeatquince🔸

What are the public's views and concerns on AI, AI ethics and AI risks?

I regulation is going to happen. A better understanding of the public's attitudes would be useful for helping EA-aligned policy advocates to ensure that the regulation designed is effective at both addressing public need and ensuring that AI development is done in a safe way.

brb243

Similarly to my comment below, if you ask the public about AI ethics and risks, they can first think about themselves, even suppressing their reasoning due to fear. One should also not bore/deter people by 'what should AI do to be nice to your friends, even those who are not' but keep an understanding of prestige and importance in answering this question.

Thus, the question could be presented as a mental challenge that can inform policy regulating safe AI development, considering that the result would be superhumanly intelligent AI with extensive productive capacity and decisionmaking power and possibly understanding and being able to influence/motivate the needs of all living beings.

Then, one can think about, in the scenario of abundance and actually being able to enact good decisions, what would an intelligent entity seek to motivate so that the needs of all living beings are catered to. Maybe the needs being cooperation and skillfully increasing others wellbeing?

weeatquince🔸

Making progress on ethics.

Sometimes I think philosophers could do better ethics work if they included surveying and working with the public as part of their tool kit . What do people actually think and how do they make trade-offs?

One specific example: I had a recent chat with a bunch of philosophers who said the standard view in philosophy is it impossible to have (or to technically formalise) a consequentialist view of justice based ethics. This confused me because in practice people do this all the time – you can find a bunch of justice based EAs and get them to make ethical trade-offs and it becomes pretty consequentialist pretty quickly (see here).

Nathan Young

How do people feel about hearing about effective charities?

How do people feel about hearing about effective careers?

weeatquince🔸

Is this question asked with the intention of maybe doing such surveys?

I do plan to do surveys of the public's view of what a good future is and would really appricate support on that. I hope to be able to fund any such work but yet to be confirmed. Would you be invested in collaborating?

brb243

First, one should ask what non-elites can do to make great positive impact. What comes to mind is donating, learning about EA, developing solutions, and presenting them to their networks. In addition, I was thinking about doing in-network elites' work so that the privileged individuals can more fully focus on EA-related advocacy within their circles.

Why one would seek to refrain from approaching the public is that 1) reputational loss risk based on a public appeals to reject EA, 2) upskilling relatively large numbers of persons whose internal professionalism standards do not reflect those of global elites in time-effective communication norms requires specialized capacity investment and 3) sharing EA concepts in depth with a large number of individuals would constrain experts in the community.

There should be people who can (2) coach relevant professional communication while maintaining openness to an individual's expression and (3) people can be encouraged to engage with more senior people only after they extensively learn on their own and with peers, so EA should have the capacity to address these two concerns.

The remaining challenge in approaching the non-elite public is (1) minimizing reputational loss from public appeals to reject EA. This can be done by avoiding individuals who would be more likely to advocate against EA and developing narratives where such public rejections would benefit the community.

Thus, some relevant questions can cover opinions on the idea of continuous pro bono learning on how to benefit others to a greater extent, perspectives on preferred learning models, linking social media posting and EA-related learning motivations, and ads that would motivate respondents' peers to start learning. Then, the appropriate ads can be offered to low reputational loss risk and high participation potential audiences based on their social media activity.

In addition to gathering data on what advertisements would invite the right people to the community, I thought of gaining the determinants of persons' wellbeing in order to identify possible win-win solutions and conducting a network analysis to target nodes of influence that have the greatest wellbeing impact.

Peter Slattery 🔸

Thanks for this, David.

Quick thoughts:

I think that an annual 'living survey' of the public could be very helpful. Maybe it could, for example, track public trends in moral views, attitudes towards activist 'brands' (e.g., EA, vegan activism, extinction rebellion), key EA organisations (e.g., FHI, 80,000 hours), key EA behavioural outcomes (e.g., supporting effective charities/caring about longtermism) could be very helpful.

Ideally I would want the survey to help the EA community to have better i) aggregations of public behaviours and attitudes (e.g., what demographics/geographies/groups do/think), ii) awareness of internal and unobservable behavioural drivers and barriers (e.g., whether people are fail to act as hoped because they are unaware, unable, or unmotivated, and why), iii) forecasts for future behaviour (e.g., whether people expect to think or act more or less optimally in future), iv) audience targeting (e.g., who we should target our outreach for best effect) and v) intervention tailoring (e.g., what to say to whom to get the best outcomes).

Knowing the combinations of experiences, beliefs, value, abilities etc that differentiate EAs from non-EAs could be very helpful, so it would be great if we could easily compare the results from the public survey against a similar EA sample (maybe the EA survey) and track divergence and convergence over time.

This would all help (a little) to answer many important questions that I, and other EA actors and organisations, seem to regularly have.

Comments