I would argue that the ability to "prove" your transaction is not hugely important. Such is the nature of cash. You probably wouldn't want to use cash to make public transactions, nor would you be inclined to use Monero.
We should focus on making XMR as untracable and unlinkable as we can (like cash), while continuing to improve usability, and worry less about these convenience features until the coin is more mature. I truly think it will be a mistake to try to make this a do-it-all, something-for-everybody currency.
For usability's sake, there should be a default ambiguity (that isn't zero. Maybe 1 or 2?). Setting specific levels of ambiguity should be an advanced or custom parameter that the average user shouldn't have to think about for common transactions.
Unfortunately it's not possible to completely disable the ability to prove a transaction is yours. After all, your "balance" (for want of a better term) is really just the sum of all your unspent inputs, determined by stepping through the blockchain with your private key and determining, for each transaction, if it contains an output or input related to you. Thus, at its worst, a user could move funds to a new wallet and reveal their private key for the old in order to expose transactions. We have to accept this as a given.
Thus, we may as well provide the functionality in a way that doesn't compromise users on either side of the transaction. BUT - you are right, it is not a major focus right now. There are stacks of other things that have to be done before we get to this tooling, but it is on the radar.
Changes to increase resistance to attacks that attempt to reduce the anonymity set are under discussion (as evidenced here, but on IRC more), so we definitely aren't letting things stay as-is:)
... please understand this is such an edge-case that we're taking about protection from an extremely sophisticated attacker with extremely large access to resources.
Such as a certain 3-letter government agency? These are exactly the people we're trying to keep out. Otherwise we'd just stick to coin-join.
Monero is at least 12 months away from anything close to this, and it's hard to determine when it is "safe enough" for this. My comment was just to point out that to the casual observers, very technical people, and even reasonably sophisticated attackers Monero is already "safe enough". To a more sophisticated attacker, especially the aforementioned 3LAs, there are so many edge cases we need to deal with that it will take time. But in its current form, it already blows CoinJoin out the water.