As we all become more and more dependant on radio signals it makes sense to keep a close "eye" on them. A bit of RDF would reveal the bogus tx.

GPS is a great convenience but it should not be a replacement for the ability to navigate. Electronically comparing the GPS to a good inertial navigation system should be able to spot bogus GPS location quickly enough to avoid problems if you have a competent navigator.

With people using autopilots and the old adage that a computer never lies that was used to be drilled into people - how many people follow satellite navigation blindly till they litterally drop off a cliff - and still say - well I was following the navigation. Always relying on the human or analgue element is a good fall back, but in the computer age is probably not the answer.

Perhaps with ships etc. it might be possible to have some direction-finding equipment that could confirm that the signals are actually coming from where the satellites are known to be (this need not be too precise because it would be impossible for a hoaxer to transmit a signal from anywhere remotely close to the proper location of the satellite.
I cannot see this working on cars as I imagine the direction-finding equipment would be quite large.

"The spoofed GPS then reported a difference between the ship's location and the desired course, and altered course to return to the correct course."
Seems to me this is where you would build/program something to relise when your being spoofed, im not sure all the ways that this would happen normaly, where your driving in a straight line, and then suddenly your pointing somewhere else without having turned at all. I mean you should be able to build hardware that monitors the actual steering and if the boat turns when it hasnt been steered then your possible being gps spoofed.
Of course i guess its possible that weather/high wind could blow you off course. they should invent something, that only works at night, that scans the night sky for stars to align to so they could get an exact position. I wonder if there is a way to do this during the day to, i mean we have telescopes that can observe stars during the day using different radiation then the visible spectrum, i think.

Seems like inertial, compass, and radio beacon data could be integrated with GPS so that anomalies could be reconciled by humans and/or heuristics.

Seems like a cross check between the propulsion system and the GPS would therefore indicate the error--the ship says it's turning, but the GPS says no. Who wins?
If landmarks are available (tough at sea, but stars at night could be a fall-back), the advertised GPS position and heading could be loaded into a digital map or star chart and compared. Any difference, who wins?
Great work. I'll have to check to make sure that this is corrected before I buy my Google Autopilot Car.
Of course, deviation of just a few degrees might go unnoticed, sufficient to drive your yacht into the jaws of a secret interceptor! (http://jamesbond.wikia.com/wiki/Liparus)

"GPS is a great convenience but it should not be a replacement for the ability to navigate".
The ability to navigate using what? Even the old radio direction finding is easily hacked. Or do you mean manually with chronometers, sextants, astronomical observations and tables? Good luck with that!

Probably the magnetic compass is the least easily spoofed device (at least not easily spoofed from a remote location (a (mechanical, or laser ring) gyrocompass must be periodically reconciled to a magnetic one, as the gyro suffers drift due to the earth moving through space (MEMS and laser ring gyros, suffer additional drift due to random processes at the smaller scales)
For a given location, and a given vector, there will only be a single magnetic course which is correct, IF there is no ocean current or prevailing wind (then the compass direction will vary according to the combination of all factors.) So it may not be so easy in nonideal conditions to tell if one was being spoofed just by looking at the other instruments.
Probably the best way to check would be if the spoofing signal correlated perfectly to the GPS almanac (direction finding would be a little more cumbersome), though I would think that the spoofers would be smart enough to include a random dither in their time signal to replicate transmission through the ionosphere (just run a ionospheric delay model).
This experiment could be inherently dangerous, if the experimental spoofing signal were received by other craft (surface or airborne) in the vicinity of the cooperating vessel..
I think that for civilian operations, navigation aids must be taken on trust, given that there are many movements of surface and air craft daily which rely on such signals. In the future, possibly anyone wanting a more spoof proof signal will have to subscribe to a secure encrypted provider, for all navigation aids, and thereby have greater confidence that the signal they are receiving is valid.... But then what happens when the provider gets hacked....
This sort of attack is little different to the lights used by ship-wreckers back in the day, just a little more technologically advanced. Who would have thought that the US defence forces would have been unprepared for this back when Iran downed the Sentinel. (then who would have thought that they weren't encrypting all of their comm's traffic.

I wouldn't be too concerned about interrupting ship navigation. Any experienced captain maneuvering in a critical situation will use multiple sources of information. But, think of all the other systems world wide that rely on the GPS clock. International finance being the first that comes to mind.