Do measurement and collapse require an actual human? 

The word “measurement” is generally used for an action of some conscious entity which gives it knowledge of the value of some quantity that it sees as observable. So, although the observing entity doesn’t have to be human (as opposed to some intelligent animal, alien, or robot), it does have to be conscious, and so not every interaction with a another system necessarily counts as a measurement.

But if a quantum system that appears to some observer be in a pure state (which happens to be a superposition of eigenstates of some observable) interacts with a specially designed (and usually much more complex) system in a mixed state, then after the interaction it may be that the effective state of the original system ends up as a statistical mixture of eigenstates of that observable. This “decoherence” process is sometimes (incorrectly) identified as “collapse” of the wave packet, but the actual collapse to a specific eigenstate only occurs in whatever passes for the mind of the observer when that mind registers the information.

Note: The fact that we can demonstrate interference effects on a photon that has passed through a macroscopic optical system of prisms and mirrors is clear evidence that the claim that every interaction with a macroscopic system induces decoherence is balderdash.

Source: (64) Alan Cooper’s answer to In quantum mechanics, does a measurement include when a particle wave comes into contact with another system even if no human is there taking a measurement? In other words, is an actual human measurement necessary to cause a collapse of the wave? – Quora