Need help with phantom audio being generated on processed files

Hello - been using the AI enhancer and it’s really been a game changer in cleaning up audio. But I have a big issue. I’m uploading a co-host voice track from a podcast recording, so it’s just her isolated voice. And in the gaps where I’m speaking and her track should be silent, the enhancer creates audio that sounds like a garbled person speaking nonsense. It’s not her voice, it’s not my voice - it’s some weirdly created artifacting or something. I have included a sample. So then I have to go through her whole track and mute those sections where this is generated, which is throughout. It doesn’t generate while she’s speaking, thankfully, just in the gaps where there should be silence.

I saw a couple of other posts about this, but no answers or resolution. Thanks!