"How disruptive will Copilot be for the Sotfware (sic) Engineering Industry?" - probably no more disruptive than blindly copy-pasting from StackOverflow: Only time will show which way software gets more bugs.

Seriously, software development eats itself for decades now. Abstractions of all kinds (both domain and solution side) become better managed, quality ontologies pop-up all the time. Thinking of AI take over software industry is similar to think genetically modified horse would win Formula 1. I believe neats shall win scruffies one more time, because if some code can be generated it may simply mean it is just commodity, and source code can actually jump to higher level of abstraction instead of being generated. To use analogy, instead of "applying AI" to do repetitious summation, inventing multiplication is the real advance.

AI can still be useful in finding patterns and help people come up with useful ontologies where a large amount of data is involved (quite similar to factor analysis, other statistical methods, or something like formal concept analysis).

Trying to make guesses what one wants to express using massive statistical data from open source code of differing quality solves the wrong problem.