To the chagrin of some diabetics using insulin, simply using an insulin pump is not just like having a pancreas. When I started getting information about my conversion to an insulin pump, I had illusions of this sort of thing being available now (well, "then"). I had half-expected my insulin pump to know what my blood sugars are and for it to compensate based on that. That is not the case, but some people/organizations say that this is not something that is too far off.
That statement must be tempered by a few precautions, however. There are organizations working towards creating a "closed loop" system, whereby a CGM (Continuous Glucose Monitor) would transit glucose levels to your insulin pump and your pump would increase or decrease the amount of insulin being pumped into your body. That seems like a relatively easy fix when you look at just that, because there are CGM's out there on the market and they do connect with insulin pumps. However there are other things to think about in that scenario. One item that quickly comes to mind is how the system would correct for things like exercise. Take the case of you taking some extra carbs before going on a run. Currently, your CGM would show you having a higher sugar level and tell the pump to give you more insulin, and when you go on your run, you now have the insulin and the exercise "taking care of" the additional carbs in your system. Therefore, a system that solely gives you insulin will never (in my mind.....non-expert, as it is) be able to be totally self-sufficient in functioning as an "artificial pancreas".
That leads to another precaution that comes to mind. A pancreas' function is not solely to create insulin. A pancreas also produces amylin (slows the digestion and the rate of glucose entering the blood stream) and glucagon (raising blood sugar). For a fully functioning "artificial pancreas" you would need this system to be also able to at least give you both insulin and glucagon, in my opinion. In a pump form, that would require at least three probes inside your body (1: the CGM; 2: the insulin cannula; 3: the glucagon cannula).
One more precaution to consider is that current CGM's (at least any I know of) only test the interstitial fluid (the dumbed down definition of which is the fluid found just under the skin) which only shows what the blood glucose levels are after a delay (studies differ on what this delay is, ranging from 5 - 20 minutes). In periods of relatively static blood glucose levels, that is not too much of a problem, but in times of sharp increases or decreases, those minutes can be very important, if you are only using the CGM as a measure of blood sugars (something you can't do).
Having said all this, there are studies currently underway in Europe and North America (FDA recently allowed for some tests in relation to an artificial pancreas) to create something that would function as an artificial pancreas. Advances are occurring all the time, and the "construction" of an artificial pancreas I have just discussed is only one way it is being looked at. I'll leave it others to come up with how to build it, but being able to have one in my lifetime would be amazing!