With all the technology that exists today, I'm surprised that we haven't invented something that would make it logistically and economically feasible to do a quick scan of e.g. all containers going into a port.
Take the Evergiven. It can fit ~20k containers. A “quick” check each going 2 minutes would add 40k minutes to loading, or 667 hours or 27 days. A month basically.
In a world where time is money no way they are checking all containers.
The hypothetical I'm imagining is that the trucks/trains going into a port go through something like a strip photography [0] x-ray machine, which doesn't need the vehicle to even stop at all. Some barcode/QR code on the side of the container connects to some manifest, and then a human (or AI?) can do a quick sanity check of "oh the manifest says this container is full of teddy bears but it sure looks like there's a car in there."
Obviously if it were that easy then somebody would have done it already, but I don't immediately see why that definitely wouldn't work. My guess is either a fast x-ray machine is implausible, and/or simply matching the contents to a manifest wouldn't be enough of a deterrent to criminals.
The real world disagrees with your conclusions. The US scans effectively all incoming containers, although they are not looking for stolen cars: https://www.dhs.gov/cargo-screening
the relevant text is near the bottom, under Results:
Cargo screening
As required by the 9/11 Act, 100 percent of all cargo transported on passenger aircraft departing U.S. airports is now screened commensurate with screening of passenger checked baggage.
International inbound air cargo is more secure than it has ever been, with 100 percent of identified high risk cargo being screened.
CBP now screens 100 percent of southbound rail shipments for illegal weapons, drugs, and cash, has expanded Unmanned Aircraft System (UAS) coverage to the entire Southwest border and completed 651 miles of fencing.
CBP has deployed Radiation Portal Monitors and other radiation detection technologies to seaports, land border ports, and mail facilities around the world. In 2003, these systems scanned only 68 percent of arriving trucks and passenger vehicles along the Northern border, no systems were deployed to the Southwest border, and only one was deployed to a seaport. Today, these systems scan 100 percent of all containerized cargo and personal vehicles arriving in the U.S. through land ports of entry, as well as over 99 percent of arriving sea containers.
But the goods going into the containers are coming from somewhere right. Why are they not checked at source or when they are due to be loaded?
If we wait until they are already in the containers, then yes, it is not feasible to check them all. Basically, the checks should be distributed, not concentrated at the port.
This would require a "container cop" at every loading point.
Containers get dropped off (empty) and loaded in all sorts of unexpected places. Not just big factories.
The big loading points would be easy to inspect, just require an independent property master or something. But you can load a container in your driveway and have it picked up for shipping to anywhere in the world.
We do have it. We just don't scan outgoing shipments. The US claims a > 99% scan rate for incoming containers, although they are not looking for stolen cars.
The OP claims many stolen cars are exported under false manifests, so identifying a car and being able to compare that to the manifest helps. Once you can use manifests to identify cars (because you can spot containers trying to hide cars) you can put them through a different process – presumably this is part of the reason the manifests lie. Another thing is that I suspect most legitimate car exports go through a few companies so it may be that other shipping details can give big updates to your prior of whether it contains a stolen car.
