More like extrapolation. Higher error bounds.
But how EXACTLY do you set the error bounds on this "synthetic instrument". Used to be that when an instrument failed, we accepted that we had lost the data it was producing, and moved on.
If this is now "science", we have a problem.