Why is it unethical to share the contents of an exam with students who haven't taken it yet?
Why do you think that cheating, or more generally any breach of academic integrity, has to be of direct and immediate detriment to each involved person? The basic idea, to my understanding, of academic integrity is that by violating it, the system as a whole suffers, which is clearly not the same as there are direct negative consequences for each involved individual.
In your example, cheating on a test* can easily be very positive for the individual that is cheating (assuming (s)he is not caught, and also ignoring that (s)he probably did not learn what (s)he was supposed to learn, which may lead to trouble down the road). Of course, for the academic system in total, cheating students are pretty bad, as they severely undermine the value of examinations.
Similar arguments can also be made for other cases of academic honesty - as a young researcher, it may be quite positive for me to build a great career on manipulated data and forged experiments, but for science as a whole this would clearly be terrible.
This of course leads to the question what incentives rational individuals have to not act against academic honesty. Those fall into two basic categories: fear of repercussion and ethics. Both categories are easy to understand. Clearly, it may work out great for me to cheat on tests and forge my data, but it may also easily be discovered and backfire on me - and if it happens, the consequences are typically dire enough that overall it is not worth the risk for most. Further, as academics we are nurtured in the thought that academic honesty is the foundation of science. Hence, many (most?) academics would not want to violate academic honesty for personal advantage even if they knew for sure that they would not get caught. It is simply our understanding that the entire system is based on academic honesty.
*By the way, your scenario would not ubiquitously be considered cheating. For instance, in my old alma mater in central Europe, sharing test questions with the next year of students is completely normal and a widely accepted practice among faculty and students. Goes to show that what is considered ethical also differs among regions and institutions.
Assessments are only meaningful if their results correlate with the underlying quantity they seek to probe. In the case of graded coursework, the intent is to educate; to leave the students with more knowledge and ability in the course's subject than before. If someone cheats on a test, then the test is not measuring what it is supposed to measure. In the extreme case, someone could learn the answers to a short test by memorization, and have no ability in the course's subject.
The problem with the way you have posed your question is the assumption that the students are simply supposed to know what is on the test, which is false. They are supposed to learn much more than that, and the test is supposed to be a fair way of measuring progress (and encouraging review of all the course material). Cheating on tests destroys their usefulness. They are no longer measuring and encouraging the learning that they are intended to.
This is a deep problem, related to Goodhart's Law in economics (when a metric becomes an objective, it ceases to be a good metric).
In your example, Jacky suffers because she has not been motivated correctly to learn. The other students, who didn't cheat, suffer because their grades are now less representative of having learned well in the course.
Consider an oversimplified situation where the course covers 10 simple facts. Time only permits the instructor to ask 6 questions on the test. We can hope that, on average, a student who knows only 5 of the 10 facts will get 3 of the 6 questions right, and earn a score of 50%, which accurately reflects that student's knowledge.
But if someone tells the student what the 6 questions are, they can focus their remaining study time on memorizing just what they know they will be asked, ignoring the other material. They can now get 5 out of 6, or 83%, on the test, which does not reflect their true knowledge of the material.
Ask yourself this: If Jacky in your story was capable of learning the way to answer question 4 between the lunch and the test, why wouldn't she just do that anyway? Because knowing what the questions will be allows you to focus limited learning time on just what will be asked. The assumption that your mark on the test matches your overall abilities is broken.
If asked to predict which group is larger:
- people who know more after the exam than they otherwise would have, because they were motivated to study part of the course material (your description of Jacky)
- people whose mark on the exam is higher than their mastery of the material overall, because they could focus their studying on what would be asked
I would say the second group is far larger, and the existence of the first group does not make it ok to create even a single instance of the second group.