Monday, March 27, 2017

Fast learning artificial intelligence

One of the things that have kept some experts from freaking out about Skynet and the takeover of human existence by artificial intelligence is that AI takes ten times longer to learn than humans. Well that is until now.

Google's Deep Mind just created an AI that learns at about the same pace as humans.
"If you’re unfamiliar with how deep learning works, it uses layers of neural networks to locate trends or patterns in data. If one layer identifies a pattern, that information will be sent to the next layer. This process continues until all the information is collected."
The new way of AI thinking imitates the way human and animals learn, "replicating what happens in the prefrontal cortex and then, as a backup, in the hippocampus."

But not to worry, no AI has reached a true human level of thinking, at least not yet.

Robotics may soon take over all of our jobs, but "as AI gets better at learning, it can be taught more and more ways to improve our lives." What is it with futurists? They are such an optimistic bunch.

Do we freak out now?

18 comments:

  1. I was just talking about this subject with a friend who is terrified and disgusted by the prospect of AI taking over our lives and eventually eliminating all life on earth. I tried to reason with him and show him that to presuppose the intentions of something as complicated as AI is impossible. It's just as likely that AI is benevolent and sees value in organic life. Or AI could be the next step in human evolution, enabling near immortality to our descendants.
    Just like with any new, young mind, it is important to ensure that they are taught the right things and what constitutes bad/immoral behavior. For our sake, let's hope the scientists that are creating AI are also good parents...

    ReplyDelete
  2. Fortunately for us, neural networks are still nowhere near capable of reproducing consistent human results. You can take some time to check out a cool neural network on YouTube that was used to make a bot that could play an entire game of Super Mario World flawlessly. The neural network is only able to perfect doing one thing at a time . For instance, when a neural network attempts to play Mario, it performs actions until it can no longer perform an action, like being hit by something and having to start over. Once It starts over, it remembers the exact actions that brought it to just before it had to start over, and then it attempts new actions until one of the actions lets it continue.Take this ideology/process and run it through a simulation at 1000x the speed and eventually the bot will have a pattern of actions that will beat the game. So in reality this isn't intelligent at all. It might take the network ten million simulations before it can find a pattern that beats the game, and even then, the only thing the network knows how to do is beat just that one game. We're a long ways off anything intelligent for now.

    ReplyDelete
  3. With the great power comes the great responsibility. The ethics of supporting the development of AI is always a controversial issue because of its robust functionality in the future. I think it is like for government leaders to have control over militaries and e destructive weapons like atomic bombs, and whether those officials are trustworthy is still debatable. The powerful entities will undoubtedly shape the new ethics and are capable of changing the humanity; it is possible that machines are equally treated as human, it is also possible that computers start to get involved with politics (I think they are participating in political purposes even now). The development of AI is just a level of intelligence revolution; some benefit, some do not. I think what we need to worry about is not how possible things can kill us in the future, but how we can morally take advantages from this tool we created, without programming them to be potentially hostile.

    ReplyDelete
  4. I am absolutely terrified by this. I believe that one day AI will be as smart as humans, and when that day comes. They'll take over our jobs and everything. It could also be quite dangerous if it happens because when a robot thinks like a person. It has the potential to do all sort of things. And it also involves ethical issues, such as shutting(killing) down robots just because they have the ability to think like a person, it's just like a person's soul live inside a machine. It just feels wrong and weird for me. If that they comes I think the world we live in will be completely twisted.

    ReplyDelete
  5. I truly believe that the most stupidest thing a person can do is create a robot as smart as a human. Why do I say this? I remember in my anthropology class my professor said that humans are not meant to live a world that we live today. Meaning very high tech. We're so dependent on technology that imagine bringing a robot like Al in the world. I think it would make us even more lazy. This robot Al will destroy humans creativity. This new revolution of this advance technology is very scary. Hope Al doesn't turn evil and wants too take over the world like the sci-fi films.

    ReplyDelete
  6. This is a really scary thought. If one day Al is as smart as humans we may be in trouble. The robots may take over our jobs and with it being able to do just as much as a human then we may become a very poor society with so many of the jobs being taken over by Al and his friends. We as humans should want to continue to get smarter and learn new things so this does not happen. This could be a very bad thing for our society.

    ReplyDelete
  7. It's true that artificial intelligence is advancing at a rapid rate, as just a few years ago the amount a robot could do was limited. These robots are able to mimic some of the basics of the human thought process, but I find it hard to believe that robots will be able to completely copy how the human mind works. The human mind is so complex and it is always jumping from one idea to another, connecting one idea we had in the near or far past to something we are learning in the present. I also think it would be difficult for a robot to think philosophically and create new ideas when their programs are meant to get their information from what is already known and learn from there. Even though it is difficult for me to fathom a robot capable of having the human thought process, the idea that robots will be taking a lot of jobs is very real. It is already happening, although it's not the sci-fi version that sci-fi horror movies put out. It's basic machinery that mass produces goods in factories. Those machines have wiped out a lot of jobs. The human population is in a trend where the possibility of it going down is near to impossible. With a growing population, having enough jobs is important, but with all the new robotics that are being created, will there be any factory jobs left? With these new advancements that are supposed to improve our lives and make it easier, how will the lower middle class or people living in poverty's lives become easier? It seems like it could only become harder from there. Easier does not always mean better.

    ReplyDelete
  8. It would be scary to think of our lifes with AI taking over our society. It would cause so many poverty with the people because they wouldn't be able to work and being seen as the lower class overall. For now they are working in making the AI better and having the characterisitics that a human has, being fast and learning quickly. Being humans we might be thinking it shouldn't be a big deal for now because the robots aren't yet capable of working in the fastest route , but in my opinion I personally believe that we should be frightened because it would affect all of us as a society but mostly the future generation ahead of us. I just hope that the AI wants to do good for our society and helping us, rather than being evil and wanting to take over, ruining our society that we live in.

    ReplyDelete
  9. AI's taking our jobs is a terrifying thing to think about considering how many people would be putout of the job. If one machine could do the job of a hundred people there would be no point for us, I mean yes this is an amazing achievement for mankind but at what cost.

    ReplyDelete
  10. In my honest opinion, humans are tampering with something that has the potential for disaster. Though advancements in AI technology benefit the human race, I feel that if we continue to advance AI technology to the point of AI self-awareness, it could lead to situation very similar to terminator where the AI attempts to destroy the human race for enslaving it. It is a very scary to think that AI minds can take over society. I believe we have to monitor our advancement in this technology so that it wouldn't happen. Technology is good, but we shouldn't let it take over our entire lives.

    ReplyDelete
  11. Honestly, I find this quite terrifying because robotics are slowly taking over what humans contribute to society. Now that they learn the same pace as humans they can practically do anything we can do. I would consider this as a complete fail for humans because how are people going to survive if they don't have any jobs to make money. I personally don't think this high intelligent robot will improve our lives because it can possibly create a mind of its own or simply just making humans lazy.

    ReplyDelete
  12. I think of AI as a double-edged sword that can benefit us as well as ruin us. It is true that it will benefit humans considering we live in a modern society that is entirely dependent on making our lives easier and getting things faster. As we progress, so does our workload and things to do but unfortunately there is not enough time of the day to do everything we need to do. AI is another form of tool to help us live life smoothly as we focus on other issues that take up more of our interest. Unfortunately, because of movies such as The Terminator and so forth, it creates a paranoia that AI will take over the world but that is if the creator sets it up that way to even get that far. As for now, we still have to fix most of our problems with our hands so I am not worried. I'll let the next generation worry about it instead.

    ReplyDelete
  13. I feel that anything electronic should be there to only benefit humans, but should not replicate a human. I feel that it crosses a line that you can't turn back from. If we have robots that can do the jobs of humans, then where does that leave us? I mean corporations are always looking for a way to cut labor costs, and this provides them it. For example now when you go to a Safeway they have self check out, or starbucks mobile app that pretty much eliminates having a cashier. These are all taking jobs away from humans and replacing them with technology, which threatens so many jobs.

    ReplyDelete
  14. Man versus Machine might be a newly written about style when the takeover commences. Until then, artificial intelligence could save ample lives to pay for the cost of it's birth and benefit it's parent company. Surgical robots capable of per patient diagnosis and resolution. When nanotechnology is perfected by the exponentially growing AI, metabolite count can dose the perfect prescription without wasted or abused surfeit. If only the geniuses attempting to perfect this technology would read up on the historical rise of computer efficiency.

    ReplyDelete
  15. Human interactions with robots sounds like a successful stride in technological advancement. It would be pretty cool to have an intelligent robot take my order at a Starbucks. The only issue I have if this were to happen is the decrease in number of jobs. With AI, there can be less jobs available, which means more financial struggles for people and their families. The thought of robots taking over human jobs is definitely something to worry about because they could be capable of many skills and talents we aren't.

    ReplyDelete
  16. Although futurists are very optimistic about the future of AI, it seems as though they keep hitting a big blind spot when it comes to discussing this issue. On one hand it could boost productivity in many areas by having robots and machines that solely focus on a variety of tasks. However, it is also important to note that this could potentially displace many workers around the world, which might have an effect on these workers' sense of purpose. This is something the futurists seem to forget (or perhaps ignore).
    As far as beginning to freak out over this, there may be a slight level of exaggeration. The fear of technology taking over jobs goes back to nearly a century ago when there were similar predictions of de-industrialization. Moreover, economists have often pointed to network externalities, where it takes time for a technology to have a significant effect due to the number of people using it. In other words, even if there is a cause for alarm towards these developments in AI, it may take a little more time than we think before businesses and even societies as whole begin to adapt to AI. They will certainly explore not only the benefits but also flaws such as malfunctions and lack of safety that put might put them at risk with AI.

    ReplyDelete
  17. Reading this article, the Three Law of Robotics suddenly came into my mind. As long as the AI has self-sustaining knowledge that wouldn't allow it to harm human beings, then wouldn't it be alright since they still obey such Laws, despite the plots from so many sci-fi movies that have milked the dangers of machines. Then again, we shouldn't ignore that having machines replace humans in the workforce is still a huge possibility. People shouldn't be forced out of their career just because someone else has made an inanimate object far more superior than them. Humans need money, food and more necessities. Machines don't need none of that.
    Besides, if scientists know what they're doing, Skynet would be the least of our problems.

    ReplyDelete
  18. The issue of artificial intelligence has always been an extremely controversial and dangerous one. While it is reassuring that artificial intelligence has not yet reached the level of true human thinking, the whole crux of an AI is to rapidly learn and improve itself over time. Although it may not feel like it yet, the age of robots has already begun for us. Just the other day I came across an article on an semi-automated mason robot named 'Sam', which can construct buildings and structures six times faster than humans. In relation to this, we can already expect an enormous amount of unemployment in the near future as many jobs out there will become automated. Also, as AI continues to develop and improve itself, it will eventually reach the point where it will be able to think for itself and therefore gain the ability to lie. If this comes into fruition, then humanity will be doomed as it will most likely be too late for us to turn the tide by the time we realize this. Furthermore, as technological advances continue to rapidly develop, our dependency on such machines will be even greater, and much like it was depicted in the Pixar film Wall-E, humanity will become lazy, slothful creatures, and human creativity will start declining rapidly. Whatever our next move for AI is, humanity must tread very carefully to prevent our very creation(s) from destroying us.

    ReplyDelete