Just fyi, Unicode keeps a official updated list of lookalike characters (www.unicode.org/Public/secur...) maybe you could submit a feature report to whatever screenreader you're using and they could make a fix. tbh I'm surprised this isn't already standard for screenreaders.
