In 1997, Deep Blue — the computer created by IBM — defeated World Chess Champion, Garry Kasparov. Some say that it was one of the first major examples of AI in action, others believe that the technology used was still quite simple and shouldn’t even be called AI.
20 years later, things are slowly starting to take shape. Last year, we heard about Google’s DeepMind beating the world champion of Go — an ancient board game known for its seemingly innumerable possibilities. While Deep Blue was originally designed specifically to play chess, DeepMind’s system has been designed to teach itself how to play the Go game (by playing with itself), and not simply carry out fixed, previously programmed moves. Deepmind’s system is general purpose and can be used for many tasks. It already has been tested with mastering many video games, for example.
For many researchers in the field of Artificial Intelligence, it was a major breakthrough, proving its potential. At this point many scientists believed that such breakthrough was still at least a decade away.
AI – The new electricity?
AI technologies are already here. When you browse Facebook or use Google to search for anything, AI algorithms try to bring you the most relevant content. More and more brands and organisations use AI chatbots to communicate with their customers. In the UK, for example,the National Health Service, offers an app that can be used by patients to diagnose their symptoms. It can recommend over-the-counter remedies and similar conventional practices. The upcoming version, will be able to diagnose over 80% of all illnesses that are generally diagnosed by primary care doctors.
Virtual personal assistants are another example of how AI is currently being used. Apple’s Siri, Google Assistant or Amazon Echo, are just a few examples. They can answer questions, book appointments, make phone calls, send messages, play music, or set up reminders. We see examples of smart devices and the expansion of the Internet of things, this area will be improving dramatically over the next few years. Recently, Facebook’s CEO Mark Zuckerberg, revealed how he programs his AI smart home system called Jarvis:
It can be used for things like turning on and off the lights, or it can tell you who is waiting at the door. Companies, like Google, are working on self driving cars that use AI technologies. Soon, taxis and cars will all be autonomous according to people like Elon Musk — CEO of Tesla Motors.
Autonomous cars have a lot of potential; they can find parking lots, potentially eliminate collisions, and even earn their owners money if rented out. In the next few years, everything will be connected to the internet; our devices, tools, homes, transportation, restaurants, stores, and whole cities — even plants — and perhaps we will also start merging our bodies with AI technologies.
How far is the future?
Some people think that we are still far away from some of these larger developments, like self driving cars or general AI, that can help us solve big problems or merging our bodies with AI. But it actually might be closer than we think. One important rule when thinking about technological developments is that these systems grow exponentially and not linearly.
What this means is that if you have, for example, linear growth and you take 30 steps, you just end up with 30 steps. But, if you have exponential growth, you take 30 steps and you get to 1bln. So, for example, 1+1 = 2, 2+2 = 4, 4+4 = 8, 8+8 = 16, 16+16 = 32 and etc. With just 30 steps, we end up with more than 1bln.
In the 1960s, a computer that cost $11 mln was so incredibly large, that it was half the size of the building at MIT. Today, our smartphones are literally a million times smaller and cheaper, and thousands of times more powerful, and this trend will continue. Everything that is currently created helps to make further improvements in the existing field, thus making things more advanced at an ever faster pace. This is why in 20 years the change will most likely be dramatic. For example, all that we use now — like smartphones, email tools, internet, cars for transporting goods, cloud systems or other ready systems — all allow us to create technologies that are faster, more advanced and at lower cost.
Experts predict that by 2025, 30% of jobs will be replaced by robots and AI systems. Just recently, it was announced that Elon Musk — the visionary behind Tesla Motors and SpaceX — is backing a brain-computer interface startup called Neuralink, that will allow for AI systems and devices to merge with human brains. The purpose will be to improve memory, or allow for more direct use of computing devices. It is claimed that the company is in the early stages, but it is known that several other companies are also working on different technologies that could turn us into cyborgs. Ray Kurzweil, known inventor and futurist, predicts that by 2030 our minds will be able to connect to the cloud thanks to tiny nanobots made of DNA. It is worth noting that researchers already used DNA nanobots to target and destroy cancer cells.
What AI means for the future of humanity?
Personally, I believe that Artificial Intelligence will have the most profound impact on humanity in the 21st century. Like many other inventions, it can be used to create good or evil. The internet, for example, provides tremendous opportunities. We can communicate with friends and family instantly, we can access knowledge about anything in a matter of seconds, but at the same time it can be used to spy on people by governments, it can be used for things like pornography, creating propaganda, etc. It can make us less connected with others in real life, as we are hooked to our devices. With AI, things jump to another level. The implications are enormous and hard to predict. AI can be programmed to self improve — Theoretically, there are no limits to what it could do and how advanced it could become.
Here are few potential scenarios:
1. General AI
General AI can be used to solve problems and answer big questions. Theoretically, it could learn anything from the internet and it could be designed to help us to find solutions to issues like intergalactic space travel or it could help to solve some scientific problems, like the relationship between relativity and quantum physics.
2. AI and longevity
AI can help us to increase our life expectancy, battle illnesses and maybe even make us immortal. Known inventor and futurist Ray Kurzweil, whose predictions have displayed an 86% accuracy rate thus far, believes that “By the 2020s we’ll start using nanobots to complete the job of the immune system,” We might be using nanobots to target diseases, or we might just use them to monitor our health. They could send us updates to our devices, so we can fight disease at very early stages. AI powered robots will perform surgeries making operations safer and more precise.
3. Merging with AI systems
One of the more controversial areas is enhancing our biology by installing electronic implants. Think about searching for something on the internet by simply thinking about it, instead of typing it on an external keyboard or giving voice commands. We already have engineers testing drones that can be controlled with your thoughts using EEG brain-computer interface:
The future might require us to install implants into our bodies, this is a controversial area, but in the same way as many women desire breast implants, many will see electronic brain implants as something desirable. And now think on how much advantage an employee who can instantly get data from the internet has over people who don’t use such technology? Your future employer can decide to hire someone with an implant instead of someone without it, which could force other people to start having them as they would fall behind.
4. Bid data and AI system optimisation
AI can help to optimise and improve the efficiency of any system. It can be used to foresee trends, find patterns and predict outcomes. For example, if you run a company that consists of human resources, financial resources, and physical and non physical assets; customers and other variables, an AI system could analyse all these different elements and be programmed to optimise performance for the entire company, find the best solutions or eliminate friction, and do whatever is required. In the future, AI will be able to run multiple operations for the company simultaneously.
5. AI and the future of the workplace
Many jobs will become obsolete. Over the next few years we will see increased demand for engineers and programmers, and declining demand for many other jobs. AI systems and robots will slowly take over human tasks. Autonomous cars will eliminate the need for drivers. AI software will be able to diagnose health problems. Amazon, for example, recently unveiled its grocery store that has no checkouts. You just tap your phone when you get inside, you pick what you like and put it in your bag. It charges you for the items you’ve picked when you leave the store. Chatbots will slowly replace customer service agents. The justice system will be dramatically transformed. Apps like DoNotPay help with legal advice. It has already helped to overturn at least 200,000 parking tickets so far. Users can provide details surrounding the violation, and the app can then just create a letter that can be used to contest the ticket. These are only a few examples, but there are very few jobs that won’t be affected by this shift.
6. The dark side of AI
In recent years some of the greatest minds alive — like Stephen Hawking, Elon Musk and Bill Gates — have all raised concerns about AI. The world becomes more and more interconnected, and in the next 20 years, most of what we interact with, will be software driven and connected through computer networks. Autonomous cars, will become aware of other cars, and work on optimising the flow of traffic to make driving safe which will all be controlled by AI.
Imagine now that there is some kind of bug or an error and suddenly all cars crash or break? Maybe someone just hacks and reprograms the system for it to crash. Maybe some group or government hacks the system, it could cause chaos. AI that, can access any system, from our homes, devices, cars to buildings and power grids, could make a big mess if used irresponsibly. In the same way that a computer virus can compromise current computer networks, it could affect AI driven Internet of things. This could have devastating implications. Financial markets could crash, every aspect of our lives could be impacted. It is certainly more dangerous than nuclear weapons.
We don’t really know how superintelligence will behave, something that is trillions of times smarter than us, could make very different decisions from us and have a different perspective on things. There is no reason to think that it will not develop strategies to secure its dominance. Imagine trying to explain to someone who lived a few thousand years ago how our world looks like today, what kind of tools we use. Can you imagine explaining things like apps and computer programs? Now think about a superior intelligence that is trillions of times smarter than people like Einstein or Tesla; the ideas that it will come up with, may be very hard for us to understand.
One thing is certain, it is coming — and whether we like it or not — we will experience some of these revolutionary changes over the next 20 years.
Singularity Or Bust
Google’s Deep Mind Explained! – Self Learning A.I.