text to speech recording sounds British when United States/English is selected

Just noticed this today.  When trying to add a text to speech recording in the Audio Library, the US English AI voices for Seraphina and Vivienne are sounding like a mix of US and UK accents.  For example, this text  "Hello! and thank you for calling!  Your call is very important to us."  The "Hello", "calling", and "call" are all pronounced with a British accent whether I choose English/United States or Auto-Detect.  Can this be corrected?  Thanks!