I came across this video that it almost indescribable in its content. What I mean is that it is funny as hell and the conversation that takes place it beyond ridiculous. The funny thing is is that the conversation takes place between two Chat Bots, which are a form, albeit a very primitive form, of artificial intelligence. If you watch the video you will understand what I am saying. This is pretty funny and felinely worth a few minutes of your time.
This is not so different from many conversations I have had with significant others. Talking in circles and the nonsensical answers to questions is hilarious, not to mention the discussion about weight.
There are so many worries, concerns, and fears that surround the invention of artificial intelligence. While I think this is certainly a scary path to travel down, I often wonder if many of the fears are overblown. I watched Chappie over the weekend, and it was awesome. It is an awesome film and one of my favorites of the year. The reason I mention this is because the movie has to do with artificial intelligence and a whole other host of ideas, and helped me realize something that could potentially happen when AI comes online.
In the movie the AI first comes online it is not very smart and does not know anything about language or pretty much anything. It needed to be taught and had to learn the way the world works. I think this is one of the key ingredients we are forgetting if and or when we come to the bridge of AI. For some reason I think we assume that it will come online and already be super intelligent and know everything and calculate thoughts with supreme logic, while it could eventually reach this juncture, I believe in the beginning it will be just like a child, which also could create a lot of issues.
There are a few problems here. The rate at which it learns and if it can be taught emotion or just logic or both. I would have to assume that it could learn at an alarming and incredible rate, so there is a fear that if it were to come online if could learn quickly and within a few weeks or even days become smarter than us. This is mostly speculation, but I would imagine that would be able to learn fast, but there is also another problem that arises. It may be able to learn fast but could it be able to understand how the world works and how to interact with it.
Let’s take the idea of communism and or socialism. While in theory it seems like a great idea and an AI might be able to understand the idea simple enough. However, it may not be able to put together that because of human nature this idea simply does not work in the real world. Another example could be something like extending unemployment benefits. The AI may think that extending them and extending them is a good idea, whereas the human will continue to take the money even though they could easily find a job. Basically they are milking the system, and if the AI were to cut them off they would have no choice but to get a job. It is almost a reverse logic that I have to wonder if a super intelligent AI would understand. This is the interesting part of AI, if it will be able to put together these abstract ideas and apply them to real life. If it can do this then it will certainly be able to problem solve at a much greater rate than we could ever imagine.
The other thing that is important to remember here is if the AI will be programmed with emotion, feelings, morals, or if it will be able to learn emotion. I think this is a crucial aspect in creating a super intelligent AI. Take the following scenario. We create an AI, it is not very smart at first but in time it begins to learn and understand our world far better than we could. It is solving problems we have as well as ones that we never even thought off. Eventually using its superior logic it comes to the conclusion that we are a plague upon the planet and the planet would be far better off without us, while this is a valid argument completely killing every human would be bad for us and morally wrong. So this is where emotion comes into play. If it can learn our emotions or be programmed with them, I believe it will be more likely to love us or at the very least want to take care of and help us. At the same time this could backfire and it could also hate us. Either way I think we stand a better chance for it to have emotion and if it can think with both its logic and at times emotion then I see no reason why it would want to harm us in any way. I am not saying that I want an emotional train wreck for an AI but I think that the more we make it like us the better we will fare. Yes we have some very horrible qualities as a species, but I would think its superior logic would prevail. It could think yes humans are bad for the planet, but killing them all is wrong, I cannot do that, but I did find a way to decrease their impact on the environment. Whereas an AI without any emotion or morals would just say yeah humans are bad I will just kill them all, problem solved.
If the AI is thinking with pure logic I believe will quickly lead us the way of the Dodo. If it can grow fond of us and have an attachment to us and us to it will only help both parties survive. Yes I realize that not all relationships end well, but this could be one of the steps that help ensure a long and prosperous journey together.
I also think that it is worth mentioning that it looks like we are far from either of these scenarios. The video attests to that but I also know that chat bots are not the pinnacle of AI technology. So who knows where we stand, depending on who you ask we are either on the cusp of figuring it out or year and years ever scratching the surface. Either way the future should be exciting.