Two disastrous aircraft crashes have revealed a problem with artificial intelligence (AI) built into the new Boeing 737 Max 8 planes. The redesign of the 737 with bigger engines created a tendency to raise the nose of the plane. The software designers added undocumented automatic controls to push the nose down under certain circumstances however the system was obviously flawed. The pilots were unaware of the automatic changes that worked against their efforts to control the aircraft and in both cases, it seems they lost the battle with the artificial intelligence.
Confidence in the artificial intelligence must have been so high that it wasn’t even documented properly. There was an assumption that this part of the program would only work to save a plane during an emergency and that it had no effect during normal operations. Clearly, they were wrong.
Boeing has announced that new software will be available next month to deal with the problem. Likely it will be found that the software did not have enough input or that it interpreted the input incorrectly to force the nose of the plane down when it wasn’t already up. How confident will they be of the new software? How will they be sure it works? Testing of the original software apparently did not uncover problems and as a result, many people died.
Do not put your complete trust in any man-made systems! While many systems work well, none of them are perfect. There is only One who is perfect and can be trusted in every situation.