In short the idea is we trust the PE's signature that everything is copacetic. In more modern fields like automotive, aerospace and software engineering that idea is laughable. Everything is about process, audit and testing.
I think most product companies are essentially self-insured. If Ford releases the Pinto and they get sued, they're paying the settlements themselves, not filing an insurance claim. So they are free to use whatever method they like to verify the competence of their employees.
That has nothing to do with the law, though. Ie. any insurance company ought to be able to decide for itself what qualifies as "engineer". There's no need for the state to define this and sue people -- the insurance company can simply refuse to sell an insurance unless the work was done by people considered, by the insurance company, as "engineers".
The biggest issue in this case is that the state was using it punatively against a person who was not providing a p eng sign off or advertising services as a p eng. Having a regulatory shared definition is fine and helps smooth commerce.
But if you want to use something that's not CSA, then it does need to be reviewed and signed off by an engineer.
A home's electrical design doesn't have to be signed off by an engineer, but it has to meet code and be inspected.
If it doesn't meet code and you want an exemption (for instance because you are doing something that wasn't anticipated by the code), then you need an engineer to sign off confirming that what you're doing is safe and truly requires an exemption.
And how do you know the process, auditing and testing was done? With documentation signed by someone who can verify this and is legally liable if it wasn't done properly.
As for legal liability, that falls primarily with the company. The chances that an individual is going to be held legally liable for a process failure in a corporation is very low, and whether or not that person has a PE will be irrelevant.
1. This person has performed the work that went into this plan/design/report or, if someone else performed it, the signing person carefully verified every important aspect of the work
2. If they used software to perform calculations, they have designed a reasonable method to check and verify the output of that software, either by using selected hand-calculations or comparing to another, independent, software system
3. They have employed a quality assurance process that is suitable for the type of work being done, which may include peer review, non destructive testing, testing to failure of a sample, etc.
4. If field work is being sealed, the PEng has personally verified the field work, or provided detailed specifications for another qualified person to do quality assurance on the field work
Now whether most PEngs really follow these "best practices" is an entirely different story. And yes, for fields that often involve very complex systems, the idea of any one single person meeting the quality management guidelines to place a professional seal on, say, an entire 787 aircraft, doesn't make sense. And when those systems also have life safety implications, it makes sense to test rigorously wherever possible.
In many fields, professional engineers are effectively delegated the authority of government under a system known as "professional reliance". Rather than the city government hiring hundreds of people to review the plans of every building, they simply require developers to have a professional engineer certify that the plans meet the building code. This strikes a balance between the city having to review everything (and the associated increase in costs), and the developer being able to tick the box saying everything is fine with no real accountability beyond their numbered company that is owned by another numbered company in the Cayman Islands.