The technology of artificial intelligence is promising, but it is not a magic potion. Oh, oh, oh, oh, Hold on. We are not against AI. By this post, we will change your mindset about what you think about AI.
Okay, let's Start...
Avoid imagining computers that can do what we can but do better when you learn about artificial intelligence
My fellow Souradeep, who has a new A.I. book, wants us to realize that the technical direction is exciting but has its drawbacks: It is less capable than people today and is coded for human prejudiciation.
I spoke to Souradeep about what artificial intelligence is, areas in which he is hoped for, and frightened of the consequences.
Soujanya: Let's begin with the fundamental elements:
What is artificial intelligence?
Souradeep: It's a term for a series of models that vaguely function like the brain for computer systems. One idea focuses on my reporting and book: a neural network which is a math machine capable of analyzing data and patterns.
For example, if you take thousands of cat pictures and put them into a neural network, you will learn to recognize what patterns a cat looks like. In the 1950s, the first neural networks were constructed, but they had never fulfilled their promise for decades. That began to change in 2010.
Neural networks have been under two important restrictions for decades: insufficient data and insufficient power in computing. The internet provided us with numerous details, and scientists finally had sufficient computer power to punch it all over.
Where will the results of neural networks be seen?
In the last 10 years, this notion has changed many technologies. Numerical assistants like Alexa, driverless vehicles, chatbots, computer systems that write poetry, surveillance systems and robots, all of them rely on neural networks to collect goods from warehouses.
Often people believe that it's like a magic potion that people talk about artificial intelligence.
Yes. Yes. Yes. The A.I. pioneers' initial sin was to call it artificial intelligence. We imagine a machine that can do everything that people can do when we hear the word. In the 1950s, this was not the case and now that is not the case.
People do not understand how difficult it is to replicate human thinking and our capacity for confusion. A vehicle driving by itself can see what it is all about — better than people can in certain respects. But it's not working well enough at all times to drive or do what you and I do to react to anything unexpected on the lane.
What are the advantages of neural and A.I. networks?
So a lot of people. The computers are able to generate huge misinformation. There's no way to say what's true and what's wrong online. The possibility of autonomous weapons is unbelievable.
And the worst is that many firms have promoted algorithms as a utopia that eliminates all human defects. This does not. It does not. Some neural networks learn from huge quantities of internet knowledge – and people have created this information. This means that we create computer systems that show human prejudice—for example, toward women and people of colour.
Some US technologists, including former chief executive Eric Schmidt from Google, claim that we risk falling behind China and that America does not take A.I. seriously enough. What is the reality of this?
It's legal, but it's difficult. Schmidt and others are trying to ensure, not just inside giant technology firms like Google, that the most significant A.I. is build in the Pentagon.
However, we must be vigilant to compete with a country such as China. Our best technological talent in the United States also comes from outside, like China. It would hurt us in the long term to close our borders to experts from this area.
So what does "TRUE AI" really mean?
What we mean today by true intelligence is whether the computer uses something called machine learning, whereby a programmer not only gives a collection of instructions that he blindly follows, but learns how to generate his own code. This is what contemporary AI computers do now. They use something called neural networks to imitate how our brain functions, in a sense. An abstract version of the human brain is a neural network, and by any definition, it is becoming real intelligence.
How Close are we to this true AI?
Machines for artificial intelligence are a long way from what human brains can do. An AI can look at a dog's image and understand what it is, but he knows not what a dog is. The definition of 'dog' does not understand. And so, even though such things can be done more quickly and more effectively and sometimes more smartly than people, the manner in which we speak about them is still not intelligent. Often we're too infected with Hollywood movies, I guess, is the problem. And so, we believe that AIs, or being intellectual, are aware of themselves as somehow. This may be in the future but we are a long way from it for several decades.
Now, a Different Question,
Is Artificial Intelligence the correct sentence?
Not in retrospect, probably. We would call it something else if we had to redefine it. Perhaps more effective machine education, deeper learning, or artificial neural networks. It's not as captivating as artificial intelligence, AI.
What can true AI do now?
The Chinese game of Go, which is regarded by people as a better predictor of creative ways than, say, Chess, by an, AI, took a move that at that time seemed stupid, but we only appreciated the genius it was in the game much later, because that was necessary as the game progressed. So she sought a solution to the problem that was not understood and not anticipated by the people that designed the computer. In these profound learning machines, this is true creativity.
It sounds amazing, but a little frightening...
Yes. Yes. Yes. This is the worry of many people. What if they appear to be more intelligent and intelligent, even faster? What if they achieve what we'd call the uniqueness – they grow 'artificial general intelligence,' become conscious of themselves, and decide not to need people. This is the Terminator films Skynet scenario. But people working in AI are much more hopeful that we are able to retain power and provide adequate security. Even if the AIs could be smarter than us, they still would be in our hands. I think we should get more upset that AI is still really stupid and that, for reasons that aren't so good, potentially stupid but influential persons regulate it. It's not good or bad technology itself. It is only the use of scientific theories. This is how we human beings use technology.
Luddites broke up machines during the Industrial Revolution because they considered them to be a challenge. Will we know the advantages of this Artificial Intelligence Revolution 100 years down the line?
That's going to be the case, I'm sure. Of course, there is often a push-back, and rightly so, at the moment when technology begins to replace human labour. People are losing their work, but AI is still creating new ones that have not yet been invented. Steam engines replaced people as they were more powerful and cheaper to carry out the work. The robotic arms on the manufacturing lines are also used. Thus, this is just another one that replaces people in those positions in a long line of technology. The difference to the AI revolution is that we really don't know how many it would have. Only for a quarter of a century did we have the Internet and the World Wide Web. If not more than the Internet, AI will transform our lives similarly and do this even faster than in the next 25 years. It comes too quickly and will change so many facets of our lives. This is very exciting, but we must be sure that we know what lies ahead. We need not to be frightened by it. We just need to be ready.
How do you think this would transform society and what benefits will it mean for us if a huge number of jobs are automated, and many things like manual labour no longer need human attention?
It could be a great thing if all work can be automated, or it can cause mass poverty, depending on what we do with all this richness. When we share it with everyone who wants it, everyone is really entitled to a paid holiday for the rest of their lives, something that many people I don't think will be against.
In fact, I think the European countries are key here because the government now provides a large number of services – and especially since WWII – in Western Europe. One would assume that because all this wealth is created by increasing automation, you only have to return a small part of that wealth to the state by means of taxes so that those who need them can provide fantastic service and can no longer get a job. Another question is how do you organize your company in order to make people feel purposeful, even without a job? It's fascinating to think about what kind of society we are trying to build, where we can prosper instead of floundering with high technology.
Why do you think the question of artificial intelligence is a significant discussion?
Since only in recent years have a significant number of leading AI researchers taken this seriously in fact in decades. Huge progress has been made in this region. If you look at those examples: when machines could beat people in Go's game? Only a few years ago, most analysts believed at least ten years would take. It happened last year. We get area by region, and stuff that people thought would take years took place much earlier, an indication of how much change is on the ground.
I think that the elephant in the room is still missing because people are talking a lot about work disturbance, mass unemployment, things such as those, but almost no scientist is talking about what will come. Machines get better and better, but at all, they get better than us, and if so, then what? Knowledge has historically been a mystery that can occur only in biological beings, in particular humans. Intelligence, from my viewpoint as a physicist, is essentially an elementary particle moving about, a certain sort of data processing. There is no rule of physics that says that in any way we can't create smarter machines than us. To me, this implies that we have seen just the tip of the intelligence iceberg and that there is this incredible possibility to activate and use all the intelligence latent in nature to help humanity thrive. In other words, I believe that most people still underestimate AI's potential completely.
How does the study of artificial intelligence relate to your physics research?
As a physicist, I don't think that human intelligence has a "magic sauce," as I said. I believe that we are a group of elementary particles assembled to help us process information in a specific way. This is also what a machine is, though. A lot of people think the robots will never be as intelligent as us because human beings are magical. As a physicist, this is not the case from my point of view.
Another way to influence my thought about becoming a physicist is that it was always said to us physicists that something was unlikely or science fiction, and then we did – it happened. If in 1915 you went and told someone about nuclear weapons, you would be dismissed as a dreamer of science fiction, who didn't know where you spoke about them. "Why do I take this seriously, if one video of the so-called nuclear explosions can't even show me?" They would say. This is absurd." This is ridiculous." However, it happened 30 years later. It would have been nice in retrospect if we had prepared a little ahead and did not end up in a very destructive race for nuclear weapons. I am more hopeful this time that if we talk about it we will really plan ahead and prevent the problems.
We still kept on learning from our errors, traditionally, through our wisdom. We fucked a lot of times when we invented less effective inventions such as fire and invented the fire extinguisher. We fucked cars up many times and invented the seatbelt. You don't want to benefit from errors of more effective technologies, such as nuclear weapons and artificial intelligence. The only time we can do it may be the only time you want to prepare for things forward and get it right first.
Do you think we really can ever know if an AI can be aware in a way we can understand?
Perhaps. In the first place, you're just food, rearranged from my viewpoint as a physicist. More simply, in a complicated way, you are a bunch of elementary particles that process information and make you smart stuff. So we know that some configurations of particles have this subjective awareness, which we call consciousness. But other configurations of quarks and electrons are also available that probably encounter nothing, like your shoe. There are also particles of the same kind. So exactly what is the difference between a conscious and an unconscious glob of particles?
You should try that for yourself if you have some hypothesis that makes the difference, as I argue in the novel. You can bring yourself into a brain scanner and have a machine that sees your brain data forecast what you really feel. Then you can compare those predictions with what you have actually encountered. If these predictions are incorrect, this hypothesis will go into the trash, while if those predictions are always right, you start really serious about this. Now the principle can not only be extended to your brain, but also to your friends' brains and computers. You take it very seriously if theory says that this machine is going through something. It is not science fiction, I think solving this puzzle is really doable. We should and should certainly do it.
Okay... So That is for Today.... We will Come again with Another article about future technology very soon.... Please comment down what you think about AI and subscribe to our newsletter for more articles like this.
To help their work, Newsmusk allows writers to use primary sources. White papers, government data, initial reporting, and interviews with industry experts are only a few examples. Where relevant, we also cite original research from other respected publishers.
Source- Science focus, 21 lessons for the 21st century by Yuval Noah Harari.