First of all check out the news story onnature.
Do you smell something fishy here? Well, I certainly do. The NASA people are planning to let the rovers decide which images are interesting and correspondingly transmit those images ONLY to earth base station. They are literally handing over the total control of selecting images to a robot. And mind you, this robot is no ordinary one. First of all, the robot is more than 78 million miles away from nearest human. Secondly, it generates its own power through solar panels.
The reason I am little bit shocked by this news is because in my view adding any form of Artificial Intelligence into objects which we cannot manipulate/observe physically is not a good thing. How can you let a mechanical thing to decide which pictures are interesting and which are not.
Then we, humans, will only receive images which rover thinks is interesting. We will never be able to see those images which aren’t interesting (from rover’s viewpoint).
Let us imagine a hypothetical situation. The rover’s AI is programmer to select images of cyclones, clouds, etc. Now, imagine if rover clicks a picture of little-green-man of Mars, its AI says that the picture isn’t a cloud nor is it a cyclone so the image is pretty uninteresting. It, sadly, does not send the picture to Earth base station. And we miss the biggest story of all time.
The sadder part is that we will [b]never[/b] know such kinds of things exist because we presume that they don’t exist. We won’t discover new things on Mars because they will be un-interesting to the AI.
This is one of the few harms of AI. It isn’t smart yet. Human intervention should always be there with AI. You never no which bug in software makes the AI self-aware 🙂
Imagine the rovers Spirit and Oppurtunity become self-aware.. Scary.. Well, this though deserves a whole thread of its own..