One of the most exciting technologies in the history of our species has to be that of artificial intelligence. The sheer, almost limitless potential for AI, machine learning and deep learning in almost every aspect of our world means that it is vitally important for us to learn as much as we can about AI and how it could affect us all in both the short and long term. In this series of articles, we’ll be looking at artificial intelligence, machine learning and deep learning and how these technologies could affect us both now and in the not-too-distant future.
The future of artificial intelligence looks set to depend on the way machine learning and deep learning are both used and developed throughout the near future. With a strong focus on both industrial and cyber-based applications for AI, many advances will likely stem from sectors such as cyber -security, financial services, manufacturing and transportation with other industries such as healthcare also being likely candidates.
With this being said, artificial intelligence technologies are already present in a wide variety of environments and enhancing the capabilities of systems and networks around the world, so what’s next? Even better machine learning? Predictive behavioral systems and AI chat bots that can not only predict our behaviors but mimic them, too? The answer is yes, almost certainly.
However, one of the reasons artificial intelligence is such an interesting area of technology is that it enables machines and algorithms to learn for themselves, enabling them to break free from the limitations of the human mind.
In this final article of our three-part series on artificial intelligence, we’ll be looking to understand what the future might hold for AI and how these technologies could go on to shape our future. Let’s start by taking a look at how AI could bring about much more advanced automation.
Near Future (1-5 years): Advanced Automation
Automation is already enhancing the way in businesses and organisations are operating and producing their products and services. With the benefits brought on by the kind of automation we already see today, enterprises and institutions are already looking for new and improved ways to enhance their automation systems. Artificial intelligence is capable of doing just that.
One of the biggest ways in which AI technologies could further improve automation is through its management of ever-larger data sets. Human beings make mistakes, this is an unfortunate fact of life and one that seemingly gets worse the more they have to do, remember, or solve. This is not the case when it comes to AI.
Given its ability to manage and interpret enormous amounts of data, artificial intelligence is now the perfect partner for automation systems and could very well be the key for other autonomous services such as predicative cyber security systems and driverless vehicles, too.
As the number of connected devices we use continues to skyrocket, the data generated by all of our connected technologies will need to be understood and then analysed and used to generate actionable results. AI-powered automation systems would not only be able to collect, interpret, and then automate certain processes and operations, they would also be able to learn from each task they were assigned, making them an invaluable tool.
Not-So-Near Future (5-15 Years): Quantum Machine Learning
Any sentence or phrase that contains the dreaded “Q” word is often met with automatic confusion or skepticism, after all, defining quantum can sometimes be slightly tricky. In this instance, quantum machine learning (QML) is a relatively recent combination of machine learning and quantum mechanics.
The idea behind QML is to look for or devise ways in which quantum software could be used to further enhance machine learning. The idea, based on the fact that quantum systems generate atypical patterns within data that classical systems are unable to efficiently produce, is that quantum computers would provide a superior alternative to classic computers when it came to machine learning.
Quantum computers are still not too well understood by the general population and remain in the eyes of many, as the Large Hadron Collider at Cern did in its first few years, as an exotic, high-concept gadget with no real-world use as of yet.
However, this notion might just be about to be turned on its head. Rigetti Computing, a California-based company whose work focuses on quantum integrated circuits, has recently demonstrated its ability to run a clustering algorithm (a machine learning technique used to organize data into similar groups) using a prototype quantum chip it had developed.
While quantum computing may still be a few years away from becoming as widespread as early AI technologies have become, it seems likely that continued progress alongside machine learning technologies as well as in other fields could see quantum computers and quantum machine learning move from physics of the far future to the technology of tomorrow a lot sooner than we may realize.
Distant Future (25+ Years): Superintelligence
Of the three future technologies to be featured in this article, AI superintelligence is both the closest and the furthest from being realized. When considering artificial superintelligence, most people may think of the AI featured in numerous TV shows and Hollywood films such as HAL 9000 from 2001: A Space Odyssey, EVA from Ex Machina, or some of the hosts in Westworld. In the real world, an artificial superintelligence would likely not feature any of the human personification the previous examples had such as human bodies or voices.
In fact, artificial superintelligences may not look like anything other than a box with a screen so as to allow it to communicate with its developers. The reason for this would initially be safety. When creating an artificial superintelligence, it is inevitable that you encounter what Nick Bostrom, author of Superintelligence: Paths, Dangers, Strategies, has dubbed the control problem. How do you control something that is so utterly superior to you?
While this may all sound like something straight out of science fiction novel, many experts believe that, with the creation of ever-more powerful machine learning and deep learning systems, what has become known as an “intelligence explosion” is probably inevitable. The ability of machines to learn and improve themselves will eventually get to the point where their cognitive capabilities dwarf that of any human being.
While nowhere close to being an artificial superintelligences, there are various game-playing computers that have reached superhuman levels of intelligence for specific games such as chess, checkers and, more recently, Go. These machines have ushered us into the era where the most proficient players of these games on this planet will never again be human. While still incredibly impressive, these AIs are only significantly good at a single task, whereas an artificial superintelligence would outperform human beings in everything.
In his book, Bostrom advises that a useful way of thinking about superintelligence is to consider human IQ levels. If the average human has an IQ of 130, an artificial superintelligence would likely have an IQ of around 6000.
With the right kinds of control systems and goals put in place however, an artificial superintelligence would quite literally be able to solve almost any problem or computation and be capable of inventing or discovering almost anything. This would completely change the way we live on Earth as such an entity would likely be able to connect to and manage the entire planet’s industrial, commercial, and personal systems, technologies, vehicles, buildings and anything else it was capable of connecting to.
Artificial superintelligence is still a very long way of, but artificial intelligence technologies have now been here for years and are still evolving and developing out in real-world environments. The Fourth Industrial revolution, as well as the Internet of Things and 5th generation (5G) wireless communications networks will also likely shape the direction of AI development in the near future. With the future seemingly destined to intertwine with artificial intelligence, it makes sense for us all to stay informed about these technologies and how they may affect us in the future.