Let's hope that a near future AI based apps "once and for all" cure for cancer won't be delayed by a temperamental app.
Does that "I cannot..." constitute free will? If so, the lid is already off Pandora's box. (That is not to say that it wasn't correct, however.)
It was probably trained on data of people refusing to help somebody. BTW If nothing fundamental changes in physics, nobody really has free will. We can only talk about human agency.
As to your first assertion, possibly, but what would that have gained its developers? As for the second, I choose. My choices may be constrained by circumstance, but I choose. That is free will. Really? More parents, or other 'agents', should tell Johnny and Suzy that they have to learn how to do things for themselves, by themselves. The wannabe coder in the article may actually come to be better at his (or her) chosen trade because of this AI saying 'no'.