A standard problem for computer security is that passwords that have to be typed on a computer keyboard are comparati...

Irina on August 12, 2019

@Marissa-Avnaim,

Let's review the argument.

Typed passwords are easy for unauthorized users to steal or guess.
A new system that relies on recognizing the voices of authorized users apparently avoids this problem.
In a small initial trial, the system never incorrectly accepted someone seeking access.
If this result can be repeated in an operational setting, then there will be a way of giving access to those people who are entitled to access and no one else.

The reasoning is flawed because:

Let's look at the answer choices:

(A) is incorrect, the argument appropriately compares a voice-recognition system to passwords that are easy to steal or guess, and concludes that it avoids this problem as it is impossible to guess and at least comparatively harder to steal one's voice;

(B) is incorrect because the conclusion is conditional. The author says "IF this result can be repeated in an operational setting, THEN..", so even though the initial trial was small, the author specifies that the conclusion is true only once the result can be repeated on a larger scale;

(C) is incorrect because it is out of scope. The argument only concerns computer security, thus any other potential applications are irrelevant;

(D) tells us the argument is flawed because it ignores the possibility that the system denies access to people who are entitled to one. Note that in describing the results of a small initial trial, the system never "incorrectly accepted someone" but what about incorrectly rejecting an authorized user? Perhaps, the system had no false positives but since we have no information about false negatives, it would be incorrect to conclude that "there will be a way of giving access to those people who are entitled to access.." as the argument leaves open the possibility that it will not, in fact, give access to all people who are entitled to it.

(E) is incorrect. Even though it is true that the conclusion is conditional, a mere requirement of reproducibility in an operational setting does not make it "heavily qualified."

Does this make sense?

Let me know if you have any further questions.