The metal detector is a tool. It can be extremely useful in archaeological research, like many of our other tools, such as shovels, old maps, or particle accelerators for radiometric dating. And like many other tools, you can also use it to destroy stuff.

The only thing that distinguishes a good metal detectorist from a bad one is how they use the detector: to add to our collective knowledge by reporting their artefacts, volunteering at archaeological excavations, or simply go out to unexplored places and discover new sites (and report them). The bad metal detectorists however go to already known sites, and take away artefacts for their own personal use, either for their private collections or to sell to others, without telling people where and how things were found (because that would threaten 'their' hunting grounds).

It is most common to have a distinction between places where artefacts are being destroyed anyway (through agriculture, or natural erosion, or whatever), and the good metal detectorist saves and recovers artefacts. The bad metal detectorist goes to places where artefacts are not threatened by destruction and does not report his/her finds. Those people are simply looters and thieves of our collective heritage (and often also of the landowners, whose permission they often do not ask!). They very well know they are doing something bad, which is why they often work at night.

In addition to what has been said, the most key aspect of archaeology is maintaining context. If we know exactly where an artefact is found, how deep it is, where it is in relation to other artefacts or features of the site, we can then put this artefact in its proper archaeological context. This can then lead us to conclusions about the site, about the daily life of the people there, and allow us to construct a more accurate image of the past.
If the artefact is just dug up and removed, however, than the object is essentially worthless. Many museums will not purchase items without knowing the context of them, and many archaeological conferences will not allow the usage of such objects in presentations, as there is too much uncertainty about said objects. This is why a lot of amateur metal detecting is bad - they could unearth something that otherwise would be incredibly useful and a good find, but by destroying its context it is rendered worthless.

As an example of the use of context, Antonio Sagona's recently published survey of Gallipoli found a number of conclusions. One of these was that the ANZACs ate only processed food, while the Turks ate only fresh food. How did he reach this conclusion? Through finding food tins exclusively on the ANZAC side of the battlefield, and finding kitchens and ovens exclusively on the Turkish side of the battlefield. Now if those tins or ovens had been removed from the site, then that conclusion would not have been able to have been found.

Also, to add to the other poster, archeology is a science, when archeologists excavate, they do it using scientific methods. For example - they "fixate" the depth of layers, possible date of certain layers, fixate the findings in relations to the layers (for future dating extrapolation in the labs). It's all part of scientific work - all that information is very important for the archeologists and historians.

The metal detectors - well, they don't do any of that, they just dig. So even if they return dug out artifact to the scientists, that artifact is pretty much useless since no one know what era is it from - what location, how and what is it related to? It's like melting ancient golden coin into gold nugget - it loses its value right there on the spot.