AI doom

 

Everybody is talking about AI and Machine learning. It is clearly on the top of the trend. Plenty of publications are about how it’s gonna help to educate people or promote goods in automated marketing campaigns. Seems like it is great invention. However, not many thoughts could be found about usage of AI on the battlefield. Just imagine, how scary it could be when machine is taking a decision to pull the trigger or not.

Time is money. Less time wasted, more money earned. Modern business is trying to be able to take decisions in or close to real time with minimum of time delay. On the battlefield we could say ‘Time is life’. Everything must be lightning fast to be step ahead of the enemy. Tardiness causes death. When marines are getting trained, the first thing they learn is the fastest watch needle because it is the most important. But that relies to human beings with their perceptions. Now when computers take over humans we could say that millisecond or even nanosecond speed of reaction is the winning bid. Military domination is not who possess more battleships, tanks and aircraft carriers, but those who could take precise and accurate decision right in time. In this context usage of ML and NN are drawing scary images.

For example, now we’ve got NN with ability to analyse images and take decision if it is a cat either a dog or a boat. With top optic power we could get images of a landscape and train network to find out if there is any noticeable changes. And if human sniper has to rely on acuity and own brain, machines in comparison can work night or day and use very powerful optic devices. Distinguishing changes NN could point directly to landscape changes and help to define a target.

Only some backward do not know what a drone is. This pilotless war machine could find any target anywhere on Earth and hit it deadly hard. However, now we’ve got operators which are sittings thousands miles away and looking at the screens to define a target to hit. It might happen, when they decide wrong and hit civilians. Now NN will be trained to distinguish who’s evil and who is not. The switch will take nanoseconds to send a missile.

Computing model could also provide some useful data to commanding officers like location of opponent’s military facilities and vehicles. Creating computing models will help to analyse logistic and infrastructure possibilities. Model will calculate several war scenarios. And all that information will help to avoid military losses and increase progress accordingly. Indeed, all that means more damage for an enemy.

So, the question is: are we releasing gin out of the bottle which could be much worse than nuclear weaponry? If having missiles stops people from making stupid decision all these years, are we not getting into the era when computers would? For the AI it would be ok, when it’ll define own casualties. ‘Well, it’s war. You have to break an egg to make an omelette’ computer will think. Hell knows what will happen in that GPU neurons. Even now we do not exactly know what is going on in these wires and boards when it comes to AI. We are living in a very interesting and exciting times. Would it be alike T2 world or digital paradise we will discover very soon.

If you wish to make a quote or have some consultation about AI or Big Data solution for your business, please contact. But it should be peaceful and moral enough.

Leave a Reply

Your email address will not be published. Required fields are marked *