One thing that made me think. While studying the subject Artificial Intelligence, which is a software or application where it can think and reason by itself and it is applied to almost everything in our lives that so called help us humans to do certain tasks. For example, now airplanes have a function called auto-pilot. It is to fly the airplane without any assistance from the pilot which is very good especially for long haul flights. I can't expect the pilot to be at helm the whole time, its tiring!
Now, before I get into it. In AI, it suppose to think like a human but of course, this cannot be done yet. Human thinking and reasoning is more spontaneous and its very complex while AI is based on facts, numbers and rules. Based on these facts and rules, this AI is suppose to help human to do certain tasks.
So back to auto-pilot, its an AI system where it helps the pilot to fly the airplane. However, after watching an episode of Air Crash Investigation in National Geographic Channel, made me wonder whether that these AI system can truly be trusted upon fully. The scenario was that one of the two airplane speed indicators was giving a false reading but the pilot continue to fly with the working speed indicator. At the time, the pilot activated the auto-pilot as they were just cruising to altitude but the pilot did not realise that the auto-pilot system is actually using the false reading from the faulty speed indicator. (1 speed indicator was at 350 knots[false] while the other is 190 knots[true] )
This is where the AI system kicks in. The rule is, if the speed is too fast, pitch the nose up(this slows the plane down). Whereby in this case, it did. However, the true speed only 190 knots, almost at stalling speed, not 350 knots. If the plane was already moving at slow speed, given that the auto-pilot making it even slower. The plane will eventually stall, hence, fall out from the sky. Killing everyone on board.
AI systems only works perfectly when everything is in good order but it is hopeless when things are not properly in order.
So, computers or human better?
1 comment:
very hard to tell. because is a mistake made by human and the AI was also written by human. So if i say the AI sucks which also means the person sucks too!
Post a Comment