It seems only one AOA sensor is currently used by software. The update will change that.
"The company said it will change the MCAS software to give the system input from more than one AOA sensor. It will limit how much MCAS can move the horizontal tail in response to an erroneous AOA signal. And when activated, the system will kick in only for one cycle, rather than multiple times."
https://www.seattletimes.com/business/boeing-aerospace/inves...
That's even worse than I thought. The data from both sensors was being logged. Some reports indicated cases of huge discrepancies (20 degrees) between sensors. It is incomprehensible to me that MCAS ignored the other sensor. The lawyers will have a field day with that and in this case I say: go for it!
Disclaimer: I've never worked on aircraft software. I've worked on unmanned spacecraft and the scheme I described is very common for fault tolerance.
Why did they keep the system on with diverging data?
It will be nice to know what are the risks when MCAS is off. Maybe x% of pilots used to the old 737 consistently stall the plane on the simulator.