How do banker bots verify transactions?

Banker bot only exchange with banker bots. Which brings up an issue, how does the banker bot company prevent fake banker bots?

Bankers issue banker bots in smart cards, and smart cards are spoof proof and counterfeit proof.  Once issues, smart cards cannot be programmed.  So, no bot appears outside of a smart card, anywhere, and banker work with the bot company to make that happen.  Hence, there is an encryption password that no one knows and it appears in every card and cannot be extracted. How does it get there? You have to have guarantees from the original bot company, there is no other way. And all ATMs, all banks, and all web sites must connected to a personalized smart card to participate. When the banker opens up the bank, he gets the bank smart cart from the vault and inserts it into the teller network. Anyone who creates a new currency on the web has to buy a dedicated smart card.

So we can see the necessity of keeping the Silicon Valley watchmakers as far from smart card as we can. The bankers association will have to keep an armed guard on the password, and the manufacturing process closely guarded. The whole system depends on banker bots trusting each other. Thus any yield curve, or probability distribution, obtained by a bot from a bot is guaranteed to be balanced to the precision of the transaction that created it.

But there are years of research in securing smart cards. So, as a service for all you who demand secure banker bots, I am reading this:

Two Factor mutual authentication with smart cards.  But that is just to educate me. Readers should know the tech community is on top of this issue, and multiple solutions are at hand.   Banker bots exchanging with each other will always be no arbitrage betters, that is, as honest to each other as the data is precise. Bankers should never have to question any bot-bot exchanges, nor should bankers ever have to worry about patents, and the leaves out SIlicon Valley, sorry.

ANd reading the paper I get this:

One remark is that in this paper, we do not make assumption on the existence of any special security features supported by the smart-cards. Instead, we simply consider a smart-card to be a memory card with an embedded micro-processor for performing required operations specified in a scheme.

Grerat, but too loose. There is no reason that smart cards can be reverse engineered any more than paper cash can be reverse engineered.