Quora – Hypothetically, if an infinitely moral entity/computer program etc had access to the whole of humanity’s data (searches, health, behaviour etc.) would it be beneficial to human kind? What would improve?
No such thing is physically possible.
Infinite morality is a myth, it implies perfect information which is not even theoretically possible.
If one assumed that one could accurately measure the quantum states and positions and momenta etc of all the particles in a single human being, there does not exist enough mass in the solar system to run such a model in real time to a degree of accuracy that would not admit of divergent outcomes.
If you use a number like pi in a calculation, it is deceptive. Writing a symbol like pi implies the existence of something, yet there exists sets of mathematical proofs that say that there is no possible integer numeric representation of pi that is perfectly accurate. One may continue to compute it for the rest of eternity.
So every time one uses pi (or any of the many irrational numbers, or any of a vast class of other classes of symbols) one must use an approximation that is useful for the context.
For many purposes 3 1/7 is close enough. For many more purposes 3.14159 is good enough. If we just restrict ourselves to the level of atoms (and ignore quarks), then in a human being we are looking about 10^28 atoms with each atom having about 10^12 collisions per second. To be able to predict where all of those atoms are going to be in 30 years time, to an accuracy of a fraction of an atom’s width, one has to use pi to a very large number of digits, and that takes a lot of computational power.
And that does not allow for the fact that many sorts of things may not be known even in theory, things like radioactive decay, photons from outside of the system (beyond the light cone of things knowable), greater than Planck accuracy position etc.
So the very idea that one can produce a computer system that has perfect knowledge is an illusion that comes from an over simplification of many different levels of fundamental uncertainty and fundamental unknowability that we have already proven beyond any shadow of reasonable doubt do seem to actually exist in this thing called reality.
All computational entities have to deal with multiple levels of uncertainty in their models and projections, and always will do. Some of those come from the nature of mathematics, some from the nature of logic, and some seem to be fundamental to the nature of this reality we find ourselves in, that we attempt to construct mathematical models of. That seems highly unlikely to ever be going to change.
Having dismissed perfect anything as myth, how about reasonable approximations?
Could we do better than we are now?
Yeah – most certainly that.
First thing, start to accept that we are fundamentally cooperative entities, and that our continued existence is predicated on that cooperation extending to all sapient entities, and realise that our existing competitive market systems are, in the presence of computational automation, now posing levels of existential risk.
We need to develop globally cooperative systems (not global control – something very different), that do actually respect the life and liberty of all individuals, and the diversity of all cultures, and do expect all individuals to act responsibly in social and ecological contexts. Liberty in this sense has necessary moral constraints and responsibilities that are demanded by a respect for the life and liberty of others. And that is a very deeply dimensional set of issues, demanding respect for diversity and a willingness at all levels to engage in reaching agreements.
Do very powerful computer systems have a significant set of roles to play in that?
Yes – most certainly that – beyond any shadow of reasonable doubt, nothing else has any significant survival probability for the vast bulk of humanity (most particularly those who currently think of themselves at “the top” and insulated from such risks).