Summary: Develop an allergy to saying “Will anyone do X?”. Instead query for more specific error signals:
– while being clear that “no” is a fine response;
– while showing willingness to change your mind so that you don’t just shut down the thread (unless you want to).
I’m surrounded by dozens of volunteer groups (e.g. in effective altruism) who are trying to help the world in one way or another, and I just keep watching them get completely destroyed by the Bystander Effect. Not the Tragedy of the Commons, mind you, wherein folks are too selfish to do something good for the group. No, the bystander effect is an even more insidious culprit in humanity’s efforts to coordinate, because it befalls even the most sincere altruists: very good people will decide not to do Important Thing X because they think someone else will do it, maybe even better than they would…
At its worst, the bystander effect is a very serious problem for humanity. For example, many hundreds of people choose not to research AI x-risk, not because they think it’s unimportant or intractable — which also happens — but because they think someone else will do a better job. (I’m estimating at least hundreds because I’ve personally heard this from more than a dozen people.)
This is too big of a problem for me to solve in one blog post. Instead, I want to start by causing my friends to solve a simple class of bystander effects occurring in email threads. Email has the advantage that it happens very frequently, so any new habit you begin developing for it will get plenty of practice and opportunities for reinforcement.
A survey of anti-bystander email tactics
I was recently emailing with a group of volunteers about putting up an awesome conversation flowhart (here’s a hi-res version for poster printing) in their workplace to advertise effective altruism and facilitate structured conversations about how to do good for the world.
Someone, let’s call her Alice, responded as follows (the underlining is mine):
|“I’ll submit a request to the X Process. Also, someone could personally ask the Y Department to post in their space; anyone want to do that?”|
The problem with this query is that, if Bob reads it and thinks he’s only the 60th-percentile-best person to do this, he’s likely to remain silent in anticipation that Charlotte — who he thinks is better suited — will respond instead. Then, Charlotte thinks the same thing, no one answers, and the thread gets forgotten about. Oops!
To avoid this, Alice could cause someone-in-particular to emit a message that will narrow the search for a volunteer. She could say:
|“… Someone could personally ask the Y Department to post in their space; Bob, any chance you’d be up for that? (no worries if not, just let us know)”|
(Yes, maybe even bold his name to grab his attention.)
The last “no worries” part is important. If you don’t do something to distinguish yourself from just trying to pressure Bob, you’ll probably feel like not sending the email, and that feeling will have a good point.
The idea is not to create social pressure to do the thing, but to be clear that you (as Alice) are waiting on Bob to narrow the search. If he says yes, the search is over; otherwise, you can ask Charlotte next.
Now, sometimes you won’t feel like asking someone for help, and you’ll realize that it’s because you’re not sufficiently convinced it’s a good use of anyone’s time. So you’re tempted to just say “will anyone do this?” to pass the buck of uncertainty.
This makes some sense, but there is a better and less ambiguous message you could emit to the group instead:
|I’m not sure conversation flow-charts are a marginally effective way for us to advertise / facilitate conversation, but if anyone else disagrees and wants to take on the task, I’ll update and support them!”|
This message helps the group to
- narrow its search for someone to put up the posters,
- update its opinion in case you’re right and folks shouldn’t spend any time on it, and
- stay curious instead of feeling shut down by your skepticism.
Another common scenario is where you (Alice) would like to find a specific volunteer, but you don’t feel competent to conduct the search. In that case, you can recurse and try finding someone else to try finding someone else:
|“… Also, someone could personally ask the Y Department to post in their space; Dave, any chance you could try finding someone to do that in the next few days? (no worries if not, just let us know)”|
The general trend I want to set here is: whenever you notice yourself creating the bystander effect, try to create a more specific and noticeable failure mode instead. You want your programs to emit error signals when they stop working, and “will anyone do X?” just isn’t a program which, when run on a group email thread, will reliably do that.
Note that in an actual room full of people, it’s much less problematic to make an open call for volunteers: you’ll know within five seconds if no one responds, and can switch to asking particular people. But emails just don’t work that way. However, you can simulate this wait-5-seconds strategy in an email thread with something like
|“…anyone want to do that in the next few days? (I’ve set a reminder for three days from now in case no one self-nominates ;)”|
Here, I’m not just recommending you set a timer, but also that you tell the group about it so it becomes common knowledge (just like a 5-second-wait in a room full of people).
I think that’s enough examples for now. To summarize, when you’re part of a small group (say, where you’re one of the 5 or 10 people most-expected to yield a response) that’s trying get something done, even if you’re not “in charge” of the thread, here are some techniques to beat the bystander effect without causing social pressure:
- Nominate specific people for specific tasks,
– while being clear that “no” is an acceptable response;
- Announce timers for open-ended volunteer calls;
- Nominate a replacement nominator if that might work better;
- Share skepticism when you don’t feel like aiding in the volunteer search,
– while showing willingness to change your mind so that you don’t just shut down the thread (unless you want to 😉
On the flip side, don’t feel paralyzed if this seems like a lot to think about. Maybe just pick one of these habits to cultivate in any given week. For me, they all mostly happen on autopilot at this point, so it definitely doesn’t have to be effortful forever. And the general attitude of querying for error signals without creating social pressure seems like enough to remember this strategy once you have the basic tactics down.