The spec seems pretty clear - it's not possible. If you feel that you don't trust them to implement the spec correctly due to incompetence or malice on their part - you need to supply the proof. If we're relying on your gut feel to be skeptical, you just look like an anti-vaxxer saying "I don't trust Big Pharma's vaccines based on my gut feel."
Even against their will, companies with capabilities can and have been compelled by government agencies to go against their promises and lie about it.
Additionally, with organizations of this size, it's very plausible that the vast majority of employees and even executives are blissfully unaware of wrongdoings, abuse and involvement in conspiracy by individual employees or decision makers.
All it takes is a single compromised or malicious person to add new capabilities in an update to an already rolled-out and well-intentioned software. Possibly a handful if you add strongly enforced and audited mechanisms for co-signing releases and whatnot.
You're making vague, unsubstantiated claims about how this could be compromised. Even if there is a bug that allows such apps to collect location info, that has nothing to do with Snowden leaks.
If a government is keen on collecting such information, they would simply not use Google/Apple's APIs and instead roll their own using bluetooth and location services. This was pointed out in the article. Now could you explain to us why you're promoting a conspiracy theory that the government will lean on Apple and Google to introduce malicious bugs into their APIs when they don't need the APIs at all?
I agree I'd trust Google over the government and many others. But the OP you are replying to called for releasing the source code, and reproducible builds. If you do that, you don't have to trust somebodies word or the laws of the land, you have the laws of mathematics in your court.
Surely you are not saying what is a clearly inferior solution is one we should be using.
Still, there are multiple plausible ways in which these apps and systems can be compromised by malicious actors (from any nationality). The risks are significantly larger than they need to, to a large degree due to lack of transparency and verifiability, and significant trust elements.
These risks are not inherent in the design of the contact-tracing system or the specific companies themselves, but they are inherent in the way smartphone apps are typically developed and distributed to end-users on mainstream platforms today.
Then on top of that, past and ongoing incidents obliterates my personal confidence that large U.S. tech companies will live up to their promises.
Given past track record of both companies, this is the only way I would ever have of not being skeptical.