Machine learning
Machine learning is where the machine part of an information system learns by feedback as it goes along.
Obviously, it depends on having a system that can adapt to the data it receives. This ranges from simple "on-the-fly" statistics to more complex things like genetic algorithms or neural networks, which mutate and compete to optimise themselves for some task.
For example, on the London Underground there are cameras linked to computers which stop too many people cramming themselves onto the platforms. These computers run neural nets, which are like a web of "nodes" between input from the cameras to the answer they come up with. First, they were "trained" by showing them vast numbers of pictures of people on platforms, letting them guess the answer and then telling them whether they were wrong or right. Over time, like rats in a maze, the computers figure out some way to get to the right response most of the time. Once they were good enough at it, they were installed in the real world and the data they receive hopefully helps them to improve by a tiny fraction every day.
Whether you think this is useful or intelligent depends.
On the plus side, it might make things safer or cheaper or save someone from a boring but stressful job. On the negative side, people are worried that they'll be replaced and that these simple feedbacks are capable of sudden insanity when the situation is not carefully controlled. Quite apart from Hal 9000 in 2001: A Space Odyssey (see fig), there are some scary stories from automatic trading and autonomous vehicles too, for example.
Perhaps more importantly in the long run, unlike proper human-type research, they don't necessarily help to explain how a particular situation works, and that makes it hard to find a way to move on. A bit like the cognitive equivalent of filter bubbles on the internet.
The {!ctf} extension on this wiki uses a broadly similar approach, but rather than restrict learning to the machine, the idea is that the feedback loop extends out to include the people using it too.
