A few years ago I read a statistics that most (90%+) computer users just do things based on rote memorization of sequences of actions, not understanding what happens in the background. I wonder how it is nowadays.
Maybe AI that can "smell" if you're on a scammy site (e.g. prompts to enter your Google/FB/Amazon password) would be useful, but then again the implementation might end up being like MS's "let us record your screen, for AI!" horseshit..
I guess it's overcomplicating things, though. Most people will just not notice that whatever domain they use is not legit.
But on the flip-side, if there's an expert on the flight, they'd notice it's a spoof site and chances are higher they'll report it to the crew. I can imagine the trick is to show a page that says "For free WiFi, follow our Instagram account, click here for our account", and the click will fail because there's no connection (if we're on the plane), but the click can trigger some Javascript to say "If you couldn't follow our Instagram, login using your Google/Amazon account: Username: ____ Password: ____". Or a more sophisticated trick would be to show a DIV that looks like Instagram and the "follow" button, which will then show the fake login...
The privacy-wary IT expert will look at the "Follow our Instagram" and think "Fuck off, I'm not doing that!" and might miss the spoofed login prompt...
If you use a random domain sure it works, but it's also fishy. I guess most people won't notice...