It seems so straight-forward from the US pilots’ reporting of similar incidents. Just disengage the autopilot and fly the aircraft. Is there something wrong with the sensor? Or are the flight dynamics such that the hardware thinks stall on a climb and the software fix is to disregard the signal based on other indicators? What will be telling will be the actions taken by the flight crew to try to solve the problem.
I would recommend one change to your statement. The problem is not the autopilot, but the MCAS, a trim adjustment system. The autopilot is not engaged until the airplane reaches 10000 feet, but the MCAS engages automatically if it senses that a stall situation exists. The crew can disengage the MCAS system using two switches that sit between the pilots. Those switches have been in the same place on the 737 since it was first introduced. Pilots and other real experts (not me) can correct my remarks.