While I can understand the desire to not interact with a user of the platform (shopify, google, PayPal, etc), because of the cost, and because many are likely to be bots or actually fraudulent. It’s precisely at these edge cases when computers and automated performs the worst.
These are the situations where the scenario is imperfectly modelled in the system. Maybe the engineers didn’t think of the edge case, or an ML model never had training data to cover it.
Point is, it doesn’t matter. People problems need people solutions. Remediating edge cases involving people’s livelihoods strikes me as most likely a people problem.
When you’re asking a person or company to trust them in a business relationship into the future, such as by being their shopfront, there’s only so many you can destroy before word gets around. I haven’t used or recommended businesses to use PayPal for a decade, because I saw them time and again inflict pain and financial stress with no recourse.
As another example, look at the general sentiment around Google Cloud. Would you build a business on top of it? Does that include an exit strategy to multi-cloud, because you’re unsure if they’ll support the service you use?
If you can’t scale to provide a human to help other humans caught in the edge cases of your own system, then you need to acknowledge you’re causing harm.
A dog food company that kills a puppy every week, is really a puppy killing company that sells dog food.
It’s a matter of perspective yes, but if you got into it to help feed dogs, surely the last thing you would want to do is harm them?
And if you became aware your killing puppies, would it not be the top priority to ensure it never happens again?