How plausible is the idea of the central banks function becoming automated. Central Banks seem to function as a feedback depending on the flow of money, inflation and fractional reserve. If the plausibility of automatizing the central bank, what kind of electronic data will it need - will it look something like blockchain technology where there is a ledger for each transaction?
It is conceivable that one of the functions of a Central Bank, of being a regulator of money supply, could ultimately be automated, if artificial intelligence were to advance enough. However, that role is much more complicated than merely feedbacks on flow of money, inflation and fractional reserve. So for the foreseeable future, that's a role filled by a panel of human experts.
And no, it will not be a distributed ledger, because that's a hopelessly inefficient way to record Central Bank transactions. The Central Bank maintains its own ledgers, and will continue to do so. It is not the case that each transaction creates a new ledger. Instead, each transaction creates corresponding entries in two ledgers - that is the essence of double-entry book-keeping.
There is an existing body of economic thought that looks at this issue - the analysis of rules-based policy. One earlier suggestion was the monetary growth rules of the Monetarists. Currently, “Taylor Rules” are a popular research topic. (There are a number of other questions about Taylor rules on this website. One that touches on similar issues is here: link to question on Taylor rules and optimal monetary policy.)
There are many variants of Taylor rules at present. Inputs can include the output gap, inflation expectations, and an estimate of the “natural rate of interest” (r*). I will just comment on the output gap. The output gap can be estimated in a few ways, and is a measure whether the economy is “hitting capacity constraints.” However, there is considerable uncertainty in the statistical estimates, and so policymakers have to end up exercising judgement about its true level.
Similar problems appear for the other inputs to the Taylor rule. This means that even if policymakers try to follow such a rule, there is discretion on what the input variables are.
Adding huge amounts of transactions to the central bank’s databases would only add limited information. What matters most is how entities view the future, and not the details of what they did today. For example, we only have limited interest in the details of the wage payments of a factory owner made today. What we are interested in is whether the factory will be expanded, or whether the owner is planning to raise prices on the output.
Finally, an important part of the central bank’s functions is regulation and lender-of-last-resort operations. These functions are critical to backstop the system, and require human judgement.