This construction appears to be dubious for it could be understood both ways: πεπεισμαι μηδένα αδικείν ανθρώπων I believe that no man does anything unjust and I believe that I do not do wrong/was not doing wrong to any man.
This is an artificially contrived ambiguity. In context (Plato’s version of Socrates’ defense speech) there’s not the slightest doubt as to the meaning. Besides, in the original sentence isn’t there a ἑκών too? That makes it absolutely clear (as if there could be any doubt) that the speaker is the subject of the infinitive, not the object. So it looks as if you’ve not only taken the sentence out of context but doctored it to make it look grammatically ambiguous when in fact it’s not.
In acc.&inf. constructions (which this is not) there are often two accusatives, one the subject of the infin and the other the object of it. In context there’s rarely any real ambiguity.
Only if ἑκὼν were accusative would it mean no-one does wrong voluntarily. That is in fact good Platonic doctrine. But grammar and context alike show that that’s not the meaning here.
Incidentally, πεπεισμαι is stronger than “I believe.”
thank you i was afraid that i had misunderstood some basics.
Constantinius, it would really help if you give the citation of what you are reading, and that you quote accurately and also provide some of the context. Greek and Latin are of such a nature that context can often make a big difference in how we read the text… 
But mwh was right I concocted this phrase based on Plato for educational reasons because my problem is that I cannot always recognize acc cum infinitivo in greek. After what verbs is it used? Verba dicendi putandi sentiendi? What else? It’s not exactly like Latin. And Smyth is too brief on the point, so maybe I have to consult a comparative grammar of Latin Greek.