If Apple claims to care about protecting privacy, the onus is on them to prove that they actually do, e.g. by publishing the source code to show everyone that it doesn't collect anything.
While I appreciate the sentiment that we should be distrustful of every large company that has our data, including apple, I take issue with the assertion that unless that publish their source code, they are selling our data.
Logically, apple has no reason to be selling our data in even close to the same way that google does. These arguments have been thoroughly outlined in the threads above. Furthermore, showing us the source code would not prove whether or not apple is selling our data, though it would certainly prove that they are collecting it. And even if they showed people their algorithms for differential privacy, you could always make the argument that they were hiding something.
Essentially it seems that there is no condition that would satisfy the criteria. It's obvious that we have to trust apple's word to some degree and that the relationship is asymmetrical. I've been convinced by the arguments about business model / risk outlined above. What hasn't convinced you?
Companies exist (only) to make money. The only reason Apple has to not sell users' data is because they think the reputational and other costs are more than the profit from selling it.
If at some point in the future, the expected costs and profits change to make it profitable, they will do it without a second thought.
If the code is closed source, the users are less likely to be able to tell whether their data is being sold. This reduces the expected costs of selling the data (in PR expenses, customers who take their business elsewhere on principle and so on).
If the code is open source, it's more likely to be more costly to sell users' data, so it's more likely that it won't be worth it even in the future.