...... I don't believe there is an inherent weakness in the design but there are undoubtedly varying degrees of quality control in the production processes used by different manufacturers.
there is also a massive difference in the service conditions that such forks see. I've seen all kinds of forks break in all kinds of ways, but not enough of any one type
personally to say for sure that X is rubbish and Y is not. I would say that in the UK, all weather/winter use greatly increases the chances of there being corrosion at the bond line, whether it be carbon to aluminium joints or aluminium to aluminium.
One of the most horrifying breakages I have heard of was in a cannondale fork, with an aluminium steerer bonded to an aluminium crown. A corrosion-assisted fatigue crack started at the rear of the steerer, (not that the owner inspected the part, but the crack would have been hidden from view within the joint), and grew until the inevitable happened which was that the fork parted company and dumped the rider in the road. The rider ended up with broken bones but survived. Others riding defective forks have not been so lucky.
Somewhat worryingly one of the most common types of bicycle recall in the USA over the last decade or so has been in carbon forks. This category includes those with aluminium steerers. How many breakages there have to be in order to trigger a recall I don't know. At least one plant in China making such forks has itself gone bust, presumably as a consequence of one recall too many.
I think it is important to note two things;
1) there is a risk/consequence thing going on here; even if there is a very low risk of failure, because the consequences are so dire, it is unwise to treat forks like other bike bits. If there is the slightest doubt about them, bin them; they can fail without warning. Also
2) it is a good idea to distinguish between failures of conception and failures of execution. The vast majority of fork breakages seem to occur due to the latter rather than the former. Bonded joints are particularly susceptible to poor QA. However it is fair to say that any fork with an aluminium steerer is likely to fatigue at some point, it is just a matter of when; aluminium is commonly believed to have no fatigue limit. The fatigue life will be greatly shortened by such commonplace things as surface damage and corrosion, which are not allowed for when the approval tests are conducted.
Bottom line; inspect your forks carefully and replace them if you have the slightest doubt about them, or they are just old/high mileage.
FWIW it is clearly possible to make a bad set of steel forks, even, but IME when steel forks fail, they exhibit clearly visible cracks and/or give due warning of imminent failure (eg by bending) rather than just snapping like a carrot. In this respect they behave differently from many other forks when they fail.
cheers