Update September 2022:
In the meantime, I have published a more in-depth look at the status quo of the field of mind enhancement, its neglectedness, tractability, and possible involvement strategies here. I have since become relatively confident that some forms of EA involvement with mind enhancement are pressing and cost-effective and am planning to ramp up field building in this space. If you think I should update towards this being too dangerous or risky, or a waste of time I would love to hear from you!
While relatively neglected within and outside EA, mind enhancement is possibly a very high impact cause area.
TLDR: This post aims to raise awareness, provide a rough framework for classification and list the most important theoretical arguments and considerations regarding the impact/desirability of mind enhancement. My aim is to kick-start some high quality theoretical discussion to gather more perspectives and takes on this. A fleshed out cause area profile of mind enhancement including more in-depth analyses of feasibility and neglectedness and a rough overview of current most promising approaches for EA is coming up.
Mind enhancement is a general umbrella term for any improvement of functions, processes and qualities related to the mind. (including reasoning, cognition, thought, language, morality, spirituality, emotions) Any intervention that achieves this will be a mind enhancement intervention, or MEI.
Furthermore, in this post, a ‘desirable’ MEI will be one that has a sustainable, long-term net positive effect on effective-altruist impact (individual and/or societal).
A lot of things fit the above definition of MEI: Examples for traditional desirable MEIs would be drinking coffee or formal education, while cocaine would probably be an example of an undesirable MEI. Both designer babies and meditation would also qualify, as well as a possible Neuralink implant or student smart drugs such as Ritalin (methylphenidate).
Topology of Mind Enhancement
Therefore, we will want to divide mind enhancement into specific areas along three basic dimensions:
Method is about the means of achieving an enhancement, broadest areas include: behavioral, pharmacological, neurotechnological, genetic
Scale is about the amount of people performing a specific enhancement, so that we can roughly differentiate effects/implications of enhancement on individual (1), group/elite (10-1000) and societal/population (Millions and more) level
Effect is about which processes/qualities of the mind the intervention affects and how.
Benevolence, Intelligence, Power framework
Since the effects of MEIs can be very broad, we will further use the Benevolence, Intelligence, Power (BIP) framework used by Rethink Priorities/Convergence for a rough characterization of enhancement effects:
Benevolence is the extent to which an actor seeks to improve the long-term future of humanity, intelligence is their ability to plan the most effective way of doing this and power is what they need to put the plan into practice successfully. (see their post for more on this)
Desirability of Mind Enhancement
Due to the scope of the subject, there are a lot of different aspects and considerations varying in relevance depending on the specific sub-areas/dimensions of mind enhancement. Therefore, final conclusions on desirability of research/increased use (and EA analysis) of specific MEIs will differ a lot. Accordingly, there is a lot of work and analysis to be done for each of these areas! A first post on pharmacological cognitive enhancement will be released soon by George Altman, a soon-to-be-doctor currently completing medical school from England. In some cases, some preliminary EA examination and theoretical research already exists for quite a while such as in the case of mind enhancement through genetic means. (EA Forum post on genetic enhancement and Shulman & Bostrom, 2003 on the subject)
In the following, I will try to give an overview of the most important theoretical considerations underlying the desirability of mind enhancement generally. Of course, applicability and relevance of these will vary across the dimensions of mind enhancement.
General Cost/Benefit Requirements
For any MEI to be desirable, it will have to meet a certain safety standard and cost-effectiveness while being sufficiently effective. In practice, this means that psychological/clinical research has to be conducted for most MEIs proving that this is the case. (especially pharmacological, neurotechnological)
There is a theoretically valid argument to be made that taking on certain health costs could be desirable if the effects are large enough. For example, it could make sense for an individual to take a pill which shortens their life-span by 10% if it increases their quality of life by 20%. However, there are some important arguments against this line of thinking. (see Coercion; Ethical, Religious or Emotional Resistance)
Mind enhancement could lead to tangible economic benefits such as decreased violence, Increased economic productivity, Improved decision-making and higher innovativeness. (See this EA Forum post for more)
Per definition, the ‘Benevolence, Intelligence, Power’ framework describes the qualities that make us care about positive long-term impact and those that let us put this thought into action as effective and large-scale as possible.
Successful enhancement of these qualities through a desirable MEI therefore means:
- fundamental enhancement of the effective-altruist impact of a given actor (when done on individual scale)
- fundamental increase of the long-term positive impact of EA (when done one group-level scale)
- fundamental increase of the EA impact of populations and society
Especially if impact is power-law distributed, targeted mind enhancement that leads to moderate relative increases in an actors’ performance could lead to disproportionally high absolute increases in their impact!
Furthermore, mind enhancement could help grow and maintain the EA community: 'Benevolent' attributes such as expansive altruism and effective-focus predict EA interest/involvement but 'benevolence' can also fluctuate and decrease over time in some individuals.
While direct effects of mind enhancement would scale linearly with the total number of individuals choosing to perform MEIs, the total bottom-line output of this could be much higher still for two reasons.
For one, the effects of improved collaboration between group members actually grow with the square of the group size, adding to the direct effects. Two, enhancement could generally improve coordination to enable larger and more efficient groups. (For more on this see Bostrom & Sandberg, 2009)
It might be impossible (Feasibility)
It might be the case, that the current MEIs in wide-spread use are the only ones that are desirable or that could feasibly work. Furthermore, if a given intervention results in an enhancement, why have we not already evolved to be that way? (a question known as the ‘Evolutionary Optimality Challenge’)
There are two ways to respond to this:
Theoretically, there are a number of reasons why evolution would choose to not enhance certain qualities of the mind, mostly boiling down to: The world we evolved to live in is totally unlike today’s world, therefore we are not well-adapted to the unique challenges of life in contemporary society. (see Bostrom & Sandberg, 2009 on this)
In practice, there are a number of MEIs that already seem to successfully enhance certain aspects of ‘Benevolence, Intelligence, Power’, such as coffee or formal education. This shows, that there must be ways of enhancing the mind and that it is plausible that there could be more.
Of course, specific sub-areas of mind enhancement will require specific examinations of their feasibility which could depend on many other specific factors. For example, the brain might be anatomically restricted which could make enhancement beyond a certain level only feasible through e.g. genetic means.
Necessity of Mind Enhancement
It is possible that some of humanity’s biggest problems are too hard for humans to solve given our current limited intelligence. Therefore, MEIs improving our intelligence / power could not only be strong contributors but even be necessary for successful solutions regarding existential risks. For example, aligning superintelligent AI might pose philosophical/technical problems which just go beyond the scope of current level human understanding. MEIs that improve human intelligence - or human collaboration - by orders of magnitude could be needed, which would likely involve technology which currently is still highly experimental or does not exist yet.
Reversal Test and Status-Quo Bias
There is a considerable amount of literature arguing against benefits of mind enhancement based on a variety of arguments. (mostly in the context of intelligence enhancement) Bostrom & Ord (2016) propose the so-called reversal test to check whether a given author’s objections to enhancement might be influenced by ‘status quo bias’, an irrational preference for maintaining the status quo.
“Reversal Test: When a proposal to change a certain parameter is thought to have bad overall consequences, consider a change to the same parameter in the opposite direction. If this is also thought to have bad overall consequences, then the onus is on those who reach these conclusions to explain why our position cannot be improved through changes to this parameter. If they are unable to do so, then we have reason to suspect that they suffer from status quo bias.”
It seems unlikely that a lot of opponents of mind enhancement would want increased use of interventions that safely lower ‘Benevolence, Intelligence, Power’. If this is the case, however, it would imply that they believe that current average levels of human ‘Benevolence, Intelligence, Power’ are optimal, which has to be sufficiently explained. This is an important consideration to keep in mind when discussing the general desirability of mind enhancement.
Enhancement of Low-Benevolence Actors
Arguably, actors with low benevolence have a disproportionately high negative impact. Hence, enhancement of ‘intelligence’ or ‘power’ of these individuals/groups could lead to bad or even disastrous consequences. This is important to keep in mind in projects looking to identify or develop new MEIs and when advocating for population scale use of MEIs which are not used widely yet.
Possible ways to avoid/ameliorate this problem are projects that specifically aim to enhance ‘Benevolence, Intelligence, Power’ in EA-aligned actors only, e.g. an ‘EA enhancement clinic’. Furthermore, projects aiming to enhance ‘benevolence’ also seem promising. (see ‘moral enhancement’)
Differential Mind Enhancement
Building on the potential risks from malevolent actors, it might prove fruitful to think about research / advocacy for differential mind enhancement: Increasing the probability that mind enhancement of ‘power’, ‘intelligence’ happens in high ‘benevolence’ individuals. And at the same time increasing likelihood of use of ‘benevolence’ enhancements generally, and ensuring most importantly that this happens specifically in individuals with very low ‘benevolence’.
One general analogy of where this is done would be background checks being performed on people purchasing fire-arms: This can be thought of as a policy that decreases the probability that malevolent individuals increase their (physical) ‘power’.
How analogous systems could be implemented in the context of the ‘Benevolence, Intelligence, Power’ of the mind might be an important area for further exploration. (see this and this EA Forum post as a good start)
Increasing use of MEIs would progressively lead to increases in average ‘Benevolence, Intelligence, Power’, eventually leaving individuals choosing to not use MEIs behind. This might be problematic in two ways: First, this could increase inequality which likely is a significant factor for societal instability and undesirable in a number of ways. Second, it leads to pressure on individuals choosing to not use MEIs, restricting personal freedom and - again - arguably leading to increased societal instability. Depending on the perceived invasiveness or ‘extremity’ of MEIs, this could be an important dynamic to keep in mind.
However, it is useful to keep in mind that this argument generally applies to most novel technologies suitable for population scale use, a recent example being the advent of smartphones who have quickly become a societal must-have leaving people not buying smartphones behind.
It might be too late
Depending on the accuracy of the ‘hinge hypothesis’ and superintelligent AI timelines, the area of mind enhancement might prove irrelevant before EA involvement can cause any meaningful enhancement, for example due to prior extinction of the human race or attainment of a hypothetical ‘singularity’. This is an argument that is applicable to most cause areas which are not directly working on existential risks, though.
Ethical, Religious or Emotional Resistance
The idea of enhancing humans through novel MEIs, even if theoretically similar to MEIs like coffee, could seem a priori off-putting to many people. This could be due to religious, ethical or philosphical views or based on intuitions such as ‘naturalness’.
This might lead to significant resistance to the concept and possibly bad press/reputational damages for the EA movement. It is therefore important to communicate that the safety, effectiveness and sustainability of MEIs is fundamental to their desirability. An additional option would be to talk of mind improvement instead of enhancement, which could have better connotations at the cost of feeling less descriptive (which is debatable, though).
Why should all this be important to EA?
As shown above, mind enhancement could be a high impact opportunity whose theoretical scale could hardly be overstated. Raising ‘Benevolence, Intelligence, Power’ through desirable means within EA or even on a societal scale could arguably be a factor to overall long-term success probability of EA or even humanity.
EA Action Points
We are still quite some time away from reaching a consensus on the ideal approaches to mind enhancement. Due to the importance of getting this right (we are theoretically literally putting the overall success of EA at stake) we will want to be sure. Therefore we need high-quality, fundamental cause area examinations to review current knowledge & identify further action points.
Furthermore more practically, any desirable enhancement of ‘Benevolence, Intelligence, Power’ must per definition be proven to be safe and effective and this is where the overwhelming majority of (mostly pharmacological, neurotechnological) MEI candidates fail: This requires psychological or clinical research which is expensive, time-consuming and difficult. EA advocacy/funding here is probably a central part of the puzzle.
It might be the case that the most promising MEI have not been discovered yet. After all, incentive structures currently lead to most medical research & developments by doctors and the industry being aimed at concrete diseases, not enhancement of the healthy status quo. Here, EA advocacy/funding strengthening fundamental research in the context of MEIs raising ‘Benevolence, Intelligence, Power’ could also be effective (despite even more uncertainty being involved)
Building on reputable data of health risks, monetary costs and magnitude of effects, something like an ‘EA enhancement clinic’ providing resources, advice and clinical guidance in the context of individual/group-level mind enhancement might be a project with very high possible impact and an important possible funding target.
- high-quality, fundamental cause area research on theoretical considerations and specific sub-areas
- evaluation of current state of the art of MEIs (safety, effects, cost, research to be done, development timescales)
- psychological or clinical research on safety / effectiveness of MEIs
- fundamental research on R&D of novel MEIs
- EA enhancement clinic
- and probably a lot more
PS: Want to collaborate working on this?
I am a medical student in Berlin very passionate about this subject, especially about the theoretical considerations and possible biomedical approaches. Currently, I am working on the fleshed-out cause area profile on mind enhancement and I am planning to then investigate biomedical approaches to moral enhancement. Hit me up if you would like to collaborate or chat/discuss about this topic. I would love to hear your take!