Unsupported. The stimulus acknowledges that computers can aid in the proof of non-computer-dependent proofs. So, it might be possible to judge that such proofs are correct even if we arrive at those proofs with the aid of computers.

Strongly supported. We’re told that human cognition can’t verify computer-dependent proofs. And if they can’t verify such proofs, that suggests they can’t detect every single error in those proofs. So, some errors may slip by that humans can’t detect.

Unsupported. The stimulus acknowledges that we can use computers for convenience for non-computer-dependent proofs. And if such computer makes an error, we can catch the error through human calculation.

Unsupported. Human ability to confirm something by calculation may mean we can be more certain about the thing, but that doesn’t imply we can be “completely” certain of it.

Unsupported. The stimulus tells us that when we use calculators to perform calculations that could be done by humans, we are not using calculators as a supplement. But that doesn’t imply it’s impossible for devices to supplement human cognition in other contexts.