The English language spreads across the globe through the internet. More non-native English speakers join the bandwagon and publish their content online. The caveat is that everyone has their own accent, and it might be a detractor for viewers and listeners across the world.
Here is an idea. It should be possible to normalize the audio to reduce the accent. First, you clone the existing audio. The cloned version would use the voice of a native speaker. The track could be generated using existing services such as Amazon Polly. Both audios could be merged to produce a milder version of the original voice.
There should be a product that does this already.