I use a card, I have limited recourse to "I didnt want to do that" about my decision to buy.

I give a store use of a card. The bank and the store have to duke it out if I say "I did not give permission to use the card like that"

I use a card online: see above. But, there's at least some protections in the scheme, if not as many as people think.

I give an AI agent located who-knows-where, under who-knows control, with who-knows KYC and AML control over the fate of my card.

* who pays if I disclaim?

* who has liability to cover the gap?

* what about card interest on the debt? The card issuer, the card bank, all kinds of people are now in the loop on deciding. Is the AI host liable at all? Why not?

I'd be super uneasy. Even in an age of 0.0003c transactions I'd be uneasy about machines spending people's money, with no clear legals about who and why can repudiate or block the transaction.

If I was a bank, I'd want the customer to have to come into my system, and set limits in ME about their spend ceiling, frequency, repudiation limits, you-name-it

Not sure they have much of a say in this. Also not really listening after seeing how they couldnt remain solvent due to issues in the last crisis - how about focusing on internal stability first before looking outward?

Reminds me of how Alexa was pitched as a tool to get people to buy more things from Amazon, but few people were willing to let it buy anything especially as Amazon's marketplace is filled with alphabet soup products and obvious fakes.

Also for many people, shopping is fun. It's not a chore to be automated away.