Because I suspect 音声は〇〇と言う is wrong.
First, 言う/言います usually sounds like it's about something habitual or something in the future. In this case, you have to use the progressive form and say 言っている or 言っています (see this).
Most of the time, you can simply omit the subject and say 〇〇と言っています. Alternatively, you can change the subject and say ～さんは〇〇と言っています (if you know the speaker) or この人は〇〇と言っています (if you don't know who is speaking).
If you want to use 音声:
- この音声は〇〇と言っています: This is not wrong, but it sounds like the focus is the deciphering/transcribing (e.g., of a noisy audio, a foreign language, a single word) rather than the message itself.
- この音声によれば、〇〇と言っています: This is natural if the message in the audio is important (e.g., an anonymous political propaganda). ～によれば means "according to ～".