Ai and Software Testing

How does AI alter your software testing strategy? Does it make testing more or less beneficial? Does it change where testing occurs in your SDLC?