Artificial intelligence is one of the most prominent technologies currently being advanced. Not only is it a hot topic for researchers, but the world’s greatest technological minds are fearful of its potential. Bill Gates, Stephen Hawking, Elon Musk, hundreds of the world’s top minds have signed papers stating their fear about the destructive potential of AI systems.
[Image Source: Pixabay]
Regardless of the top minds in opposition, advances in the industry continue. Integrated AI systems today are already helping us get through daily life, according to Wired. Siri, Alexa, and all the other virtual assistants in our world are just the tip of the iceberg to what the future may hold. It’s an exciting time in the world of programming, in technology, and in simply just in the universe. AI could lead to an age of rapid exploration and discovery – it could also lead to an age reminiscent of the Terminator series.
AI Advances Today
As we seek to understand the future of artificial intelligence, let’s begin to understand the current advances in the industry today by taking a rather extensive look.
AI Mastered the World’s Hardest Game
In January of 2016, an AI system beat the world’s best player at the game Go. If you’ve never heard of the game Go, then you likely don’t grasp the magnitude of this achievement. There are more potential moves in the game than there are atoms in the entire universe. This is a game that is so complex, that the world’s top minds thought that only humans would ever be able to master it – but now a computer has. The computer wasn’t programmed to beat the game either, it was programmed to learn how to beat the game. Through positive reinforcement from the programmers, it taught itself how to win. The video below will give you some background into the game and how the computer did it.
Beating a Go expert is one of AI’s most profound achievement of the century. It can be likened to when a computer first mastered the game of chess. Playing games is one thing, saving human life is another.
Artificial Intelligence Saving Lives
We can’t talk about AI without discussing Tesla’s autopilot system. While the system has been highly controversial in the media, it is no doubt saving millions of drivers’ lives indirectly and even some directly. According to a report from the US National Safety Council, the death rate for driving is 1.3 deaths per 100 million miles driven. For autonomous systems, they have currently driven 130 million miles with only one confirmed fatality. This presents an improvement over regular driving, but more data will need to be gathered to understand by just how much.
In a much more direct sense, Tesla’s autopilot system is credited with saving a man’s life, according to Tech Republic. A man was driving home from work in Springfield, Missouri when he began having a constriction in his chest. Tesla’s autopilot system helped him get almost entirely to the hospital. Joshua Neally, the man whose life was saved, credits the AI system with saving his life.
Predicting the US Election
The last US election cycle was one of the craziest in recent history. While most in the media predicted a Clinton win, there was one AI system that predicted Trump would take the presidency. This AI system is called, MogIA, has predicted the last four elections correctly, according to CNBC. The system was created in 2004 by Sanjiv Rai. It has gotten smarter progressively over the last decade and a half, progressively analyzing more complexities in social media engagement data.
This AI system essentially gathers all engagement data from across the web and compiles it. By doing this, the AI system can understand true voter sentiment without the mask many put up during election time. It predicts the election based on engagement numbers and it hasn’t been wrong yet.
The Growth of AI in Industry
Stepping back from the specific view of AI projects that have advanced the state of technology, we can look at how AI is evolving as a trend. According to Bloomberg Technology, Google has ramped up the number of projects it undertakes involving AI in recent years by massive numbers. In 2012, they only had about 100 projects involving AI, in 2015, that number was just over 2,700.
Technology industry giants aren’t the only ones who are adopting artificial intelligence, technology startups are basing business models around the capability. Prominently, tech startups are being founded with the sole purpose of using AI to collect big data and analytics for companies. Startups are also using AI to create widely available personal assistants like X.AI. AI programs can even write news articles now thanks to a few companies pioneering the field.
In some senses, there isn’t a piece of new technology today that wasn’t created or influenced by AI – and the field is only in its relative adolescence.
Technology Driving AI Innovation
Before we dig into the complexities of the technology behind AI, I want to prime you with some visual media.
First, a look at what quantum computing could do to AI from Google.
And next, a look into whether AI machines could deserve rights as the technology advances.
We are posed with a technological aspect to AI development in quantum computing that we’ve never seen before and an ethical problem of robot rights through refinement of AI ability. This ethical problem is something we need to keep in mind moving through technology. Regardless of your current opinion, future generations will need to answer at what point technological innovation gains its own rights respectively.
With some background out of the way, let’s dive into what is quite possibly the most exciting aspect of AI technology: Quantum Computing.
To explain it simply, quantum computing presents the opportunity for a computer to function with bits in a state of superposition. In other words, computers wouldn’t be limited to the 1 or 0 states, but they could exist in both states simultaneously. The dilemma with practical quantum computing is that while superposition can exist on a quantum level, as soon as we observe the state of the data, or read it, it is seen in either a 1 or 0 state. So, what companies like Google, Intel, and NASA have done is create quantum computers that can operate in superposition to perform operations, then convert the highly complex data from a quantum read level into a binary read level. This allows for complex operations to be completed at rapid speeds while still outputting answers in readable data structures.
A Qubit Mechanical Resonator [Image Source: Wikipedia]
According to Science Alert, Google’s quantum computer is 100 million times faster than your computer at home. This is a big deal for AI.
While artificial intelligence programs have been developed without the use of quantum computing, added super processing power will greatly accelerate AI’s expanded function. Google and other top technology companies are greatly interested in applying quantum computing to machine learning applications. They state that:
“This is because many tasks in these areas rely on solving hard optimization problems or performing efficient sampling.” ~ Research.Google
Solving problems of optimization is exactly what AI is. It’s also exactly what quantum computers are good at. The two respective industries were almost made for each other.
Cloud computing has exploded in recent years and its capabilities have significantly impacted the technology industry. You have huge industry leading companies like Autodesk switching to cloud computing for all consumers, Google making supercomputing infrastructure available to anyone, cloud presents a great computing power to the everyday worker. The following video will give you a very basic background into cloud computing and how it advances software and technology.
On a level more attune to artificial intelligence, cloud computing can supply the processing power needed to run AI programs. It is very likely that for the beginning large AI applications yet to be developed, processing speed and power will be the limiting factor. What cloud computing does is make computing power a service, not a product.
In essence, adopting cloud computing allows one central and specialized company to undertake the hardware services of the computing industry. They manage all of the equipment, the upgrades, the storage, and all you have to do is pay a small fee to get access to some computing power. The cloud keeps the small guys, or even the big guys, from having to purchase and maintain hardware and allows them access to unimaginable power.
The cloud could not only facilitate expansive AI programs to run on rather mediocre devices but it will open the playing field to who can advance the state of artificial intelligence, according to CIO. When everyone has access to the computing power needed to drive innovation, technological minds aren’t limited by the company they work for or any selective opportunities provided to them.
In summary, cloud infrastructure being laid right now is the framework for future AI programs.
One of the biggest misconceptions about artificial intelligence is that it will never be smarter than us because we created it. The problem is, we already aren’t creating AI, and the AI programs of the future will only be flutters of human inspired design. Generative algorithms are behind this, and they are changing everything about how we write programs.
Generative design is perhaps a more exciting technology than AI. It will be a technological capability that allows the rapid advance of AI knowledge in a very short amount of time. A generative algorithm is exactly what it sounds like, it is an algorithm that is programmed to generate code and create programs. A programmer, in theory, doesn’t have to write the millions of lines of code that may be needed to create an AI system. They only have to create the relatively shorter generative algorithms that will write the code for the AI program. This isn’t some far-off technology either, generative algorithms are already being put to very practical use.
Generative algorithms are designing highly technical parts for machines, they are designing art, writing music, they are creating things on part with pure creative human expression, according to open AI. This is a hard concept to grasp, but technology has enabled generatively programmed creative expression. Down this avenue, you can begin to see how robot rights may become an issue in the future.
Fears Surrounding AI
Now that we have begun to understand some of the accelerators that are driving artificial intelligence, we need to understand the aura around the technology given off by technological leaders. To quote myself here from another article examining this issue,
Stephen Hawking, Elon Musk, Bill Gates and about 100 other leading scientists and engineers believe that artificial intelligence could be more dangerous than nuclear weapons. Elon Musk specifically also believes that AI is the biggest “existential threat” to humanity’s existence ever. The world’s top minds have some pretty strong opinions about the potential that artificial intelligence holds.
While these minds have these opinions, they aren’t necessarily stressing that we should avoid AI altogether, but rather that we need to be really, really careful.
These top minds worry that AI will advance so far that we will allow it to make decisions for us, according to Wired. Stephen Hawking is very concerned that humanity will give AI too much freedom, but he isn’t concerned that AI will be all bad.
This idea of giving AI too much freedom stems back to the main concern from the tech industry. Leaders are worried about irresponsible handling and understanding of what will be the most powerful technology humans have ever created.
What’s Ahead for the AI Industry
It’s hard to wade through the predictions for AI and separate hype from actual probability. The future for the industry is no doubt bright, and it is set to revolutionize how many of our daily processes are undertaken.
China is currently looking to expand into AI research and technology, rather than simply being a producer and copier of western innovation, according to MIT Technology Review. This would allow for an even more accelerated pace towards future implementation of AI into everyday life. With this implementation, many believe that language learning algorithms will be at the forefront of commercial AI.
The ability to use a phone or an earpiece to translate spoken language in real time would be one of the most powerful achievements in recent human history. Waverly Labs has already done this with their Pilot earpiece system, but it still has some kinks to work out. AI implementation would be significant to language translation because it could do all of the fine processing needed around spoken language rapidly. In essence, an AI program would be able to listen to someone talking in one language and spit back the words in your chosen language. It would be able to avoid the somewhat funny literal translation of many translation programs and handle contextual references and understandings.
Industry leaders believe positively reinforced machine learning will also play a bigger role in future development. This is much like how the AlphaGo robot learned to beat the game Go. Programmers and engineers are expected to improve how we positively reinforce AI programs, much like how we teach kids. Improving the feedback loop would give even simple AI programs to learn complex things.
Lastly, leaders predict that AI’s ability to predict the future will significantly improve in the coming years. We have already seen glimpses of this in the AI program that predicted the US election. Circling back to positive feedback, as it gets more elections right, it will refine its algorithms to near perfection. Predictive algorithms are already being used in autopilot systems for cars to prevent accidents. AI systems can predict the future by communicating with us, thinking just a step ahead to imagine what we want next. These processes and abilities are expected to be refined to a point that, frankly, will seem like pure magic.
Artificial intelligence is both something to be direly afraid of and something to be massively excited about in the years to come. AI and its respective technologies are going to bring in the robot age of technological achievement. Exciting times are ahead.