Would You Accept a Third-Party AI We All Recognize Deciding Whether Mutual Aid Is Granted?

Could AI bring a new possibility to online mutual aid:

from now on, whether a member receives mutual aid would no longer be decided by staff, but concluded directly by AI according to the mutual-aid covenant.

The idea is actually simple, and easy to implement.

Every mutual-aid group has its own “Mutual-Aid Covenant,” which sets out who may join, how long the waiting period is, under what circumstances mutual aid may be obtained and under what circumstances it may not.

When a member applies for mutual aid, we submit their application materials, medical records and other necessary documents, together with the covenant in force at the time, to an AI large model that we all recognize.

There is only one thing the AI needs to do:

Understand the facts, then compare them against the covenant, and determine whether this application meets the conditions for mutual aid.

If it qualifies, mutual aid is granted; if not, it is not — and it tells everyone which clause of the covenant, which facts, and why it reached that conclusion.

What really interests me about this is not how much labor cost AI can save.

It is that it could change a certain way of governing online mutual aid.

In the past, whether a person received mutual aid always ultimately depended on human judgment. And wherever humans judge, members may have doubts: is it a difference in how the staff read the rules? Is personal favoritism involved? Does the platform refuse when the amount is too large? How do you guarantee that people never make mistakes?

If AI can do this work maturely, it may gradually remove those human factors.

AI does not know the applicant, does not know who has connections and who does not, and has no reason to change its judgment just because a mutual-aid payout is large.

It has only two things to face: the facts on one side, the covenant on the other. And then it reaches a conclusion according to the covenant.

What truly decides whether a person receives mutual aid should be the mutual-aid covenant that everyone recognizes.

AI is simply the best possible executor of the covenant.

Of course, whether AI might misjudge, whether different AIs might reach different conclusions, and what to do when a dispute arises — these are all problems that must be solved when this is actually implemented.

I think it is worth trying seriously. And technically it is no obstacle at all — it could even be put into practice right away.

If one day, when you apply for mutual aid, what you ultimately face is not a member of staff but a third-party AI that recognizes only the covenant and our prior agreements — would you be willing to hand that judgment to it?

This article launched a poll: as a member, do you accept the AI’s verdict?

Options: Accept / Do not accept / Not sure

Go to the original WeChat article to vote ›