AI is taking work away from contract developers, and those of us in the field can feel it.
Yet if companies bring development in-house and hand it to AI without software engineering experience, they end up entrusting product quality to AI as well.

AI extends its changes to code no one asked it to touch, then justifies them with plausible-sounding reasons.
Who can give that output the all-clear for quality?

Generation is probabilistic, which is exactly why verification must be deterministic.
How far can “impact-scope tags,” which link requirements to code, keep AI from changing too much?
We will candidly share both the promise and the limits revealed by our own experiments.

What will system integrators be chosen for? The answer awaits you at the session.