Background and what to expect. No answers given away.
The subfields are defined by what they consume
The cleanest way to keep these straight is to ask what kind of input each one deals with. One branch is about text and speech, the meaning behind words and the ability to respond in kind. Another is about pixels, identifying what appears in an image or a video feed. A third, broader one is simply about learning patterns from examples rather than being explicitly programmed.
Once you sort them by input rather than by application, the questions become much easier, because most real products combine several and the question is asking which one is doing the specific job described.
Generating is different from recognising
For most of this field's history the goal was classification: is this a cat, is this spam, is this fraud. The recent shift is towards producing new material, whether that is prose, images, audio or code.
That distinction has its own term and it is the single most important piece of vocabulary added to the subject in the last few years.
It is already in things you do not think of as AI
Shopping suggestions, the order of a social feed, spam filtering, map routing, face unlock, voice dictation: all of these are the same underlying technologies, deployed quietly enough that they read as ordinary software.
Several questions describe an everyday feature and ask which technique makes it possible. Working backwards from the input usually gets you there.
The assistants belong to companies
The named products in this space are each backed by one of a small number of large technology firms, and knowing which belongs to whom is a fair question given how prominently they are marketed.
One of them is aimed specifically at helping people write software, which is a different purpose from a general conversational assistant.
The abbreviations are most of the difficulty
This field compresses nearly everything into three-letter forms, and the questions take advantage of that. When you are asked to expand one, the alternatives on offer are not random: each is built from words that genuinely belong to computing and arranged into a phrase that sounds entirely reasonable.
The defence is to work out what the technique actually does before you look at the options, then find the expansion that describes that. Choosing by which phrase sounds most technical is precisely the failure mode these questions are constructed around.
The options reach outside the field
The alternatives you will be choosing between are not all drawn from this subject. Several belong to other parts of computing entirely, areas that are serious and well established in their own right and simply happen to appear in the same headlines and the same job listings as this one.
That makes the useful question a definitional one rather than a technical one. For each option, ask whether it describes a system that learns, interprets or produces something, or whether it describes infrastructure, storage, security or a way of representing data. The second kind is answering a different question than the one being asked.
