Paul McNeil discusses how far AI has come, with industries already investing heavily to take advantage of the range of applications.
PREDICTING HOW artificial intelligence technology will evolve in the following ten or 20 years, or even beyond, is very difficult to say the least. However, certain is the fact that there is much to be gained to go around for everyone. It is estimated that by the year 2018, robots will literally be supervising more than three million of us at work; and by 2020, smart machines will become a major investment priority amongst at least 30% of all CIOs.
Right now, many different fields, spanning from customer service to journalism, are already being set aside by increasingly able AI that can replicate human abilities and experience. Already before our eyes is an aspect that we once thought only belonged in future technology. How will it be implemented in the global market is the question before us today.
As time passes by, it is becoming increasingly obvious that industries already taking on the advantages of AI and enhancing technology as they move forward, will ultimately make this all the more useful and robust with a range of applications that is growing by the day. A momentum is being created by those organizations that can actually afford to heavily invest in AI – such as smart cities – and this is paving the path for others to follow suit. Those entities who have remained out in the cold on this are at risk of falling behind.
Comparing the risks and rewards
There are those who will argue that predicting the risks of AI applications is impossible. Moreover, these critics will say one can never determine the exact risks and rewards, let alone compare them in this field, as analysts are releasing their forecast until 2020. There are experts who believe autonomous software agents will be handling 5% of all economic transactions by 2020.
Companies willing to risk and invest in this field are in control of the future of AI. They are taking on the challenge to conduct research and push forward the needed development. There are also cases of accidental advancements, similar to the case where a company had a programmer hired for over six years and paid him $500,000 through the course, only to learn he had completely automated his job all along.
The military is a pioneer in many of the AI advancements made. The U.S. government is also a major investor, requesting $4.6 billion to fund its drone arsenal for 2017 alone. The drones currently manned and the problems evolving in this regard, have driven the industry to begin replacing them with automated drones. Just provide a destination to an AI drone and sit back and watch how it dodges all air defense thrown against it, reaches its designated destination and still allow human eyes to make the final lethal decisions, or cancel the mission it altogether.
Looking at the issue from an academic point of view, the Massachusetts Institute of Technology and the University of Oxford are two institutions working hard to map the entire human brain and go on in attempting to actually outdo it. There are two paths to tackle such a challenging objective: developing an AI that has to replicate human brain complexities, or emulating one from scratch. Here is where a long slew of ethical concerns and questions are raised, including: what are the rights of an AI device? And what if an AI emulating your loved one becomes dysfunctional as its server(s) crash or shut down?
As these legitimate questions remain a difficult topic to answer, through the span of time the benefits of AI systems – already proven for many industries – will encourage major companies and other entities from all branches of the economy to engage this developing trend. As the information technology has become indispensable to nearly all existing industries, rest assured that AI will grow at probably an even faster rate.
AI in Australia
In Australia, estimates show 40% of all jobs will become automated by the year of 2025. That is alarming to some, while seen as very profitable to others. More significant is another figure forecasting Australian enterprises poised to invest over $50 billion in AI technology in the next decade to come. The importance of this matter has gained much attention as the Artificial Intelligence for Enterprise conference was held in mid-September, shedding light on how entities and companies can take advantage of AI technology to develop innovation, business growth and strategic direction. Key for the future of the AI industry in Australia is comprehending the potential in this corner of the globe, how to gain maximum profit and value through AI implementation, and gaining an operational perspective with the goal of implementing and managing AI.
Australia's first self-drive car was unveiled in Victoria this week. The car was a joint venture with German multinational Bosch and the Victorian Government, which invested $1.2 million in the project.
The road ahead in computation
AI, up to now, has been focused on specifically blueprinting programming tools for certain functions only, proving to be extremely rigid in practice. Such computing strategies based on AI have become very common in the past few years. However, true learning is the aspect that AI has to engage in to develop into the future. For instance, AI will not have any reliance on receiving direct commands to comprehend what is being demanded from it.
We are currently using GPS systems that rely on automated learning and perception. These are mobile devices that are able to even interpret speech and search engines that learn and adapt its abilities after figuring out how to interpret our intentions. The next step in AI are seen in developments made by Google and IBM in DeepMind and Watson, respectively.
No knowledge at all was pre-programmed into DeepMind, as there exists no handcrafted programs or any modules specifically missioned for certain tasks. Learning automatically is the very philosophy of DeepMind, being crafted specifically to be general to guarantee the end result will enjoy properties of emergent nature. This includes program software capability that can take on and actually defeat grandmaster-level Go players, proving to be a very impressive show of incalculable importance when we learn DeepMind was not programmed for such a specific task.
Traditional AI can be described as narrow, only able to carry out the specific task it is programmed to understand. Olli, on the other hand, is a Watson-powered automated car that learns from first monitoring its passengers and then interacting accordingly. Whenever a new passenger requests to provide a suggestion or seek a destination, the Watson inside begins to store the data to be used when engaging the next person. New sensors are added continuously and, like a human driver, the vehicle constantly gains intelligence as it carries out its tasks.
However, Google and others are investing in these AI systems to forecast end user buying habits with more efficiency than current recommendation software. The question is: can they? Can AI dynamically optimize supply chain transactions by comparing to previous patterns? This is where real profits can be made, and this is a far more challenging and complex dilemma than playing any type of games, driving a vehicle or carrying out repetitive tasks over and over again.
Proof points provided currently by a variety of AI platforms – including distinguishing fashion mistakes or being able to predict health difficulties – are a clear indication that AI technology is expanding and, in the very near future, these tasks we thought to be complicated will become a simple reality.
Some may find it hard to digest, but in the not so distant future AI will begin mimicking complex process of human decision making, including issuing prescription to patients and suggesting investment advice. As improvements are continuously made in true learning, positions of first-tier support and jobs of a more dangerous nature (such as driving large trucks) will be carried out by robots with full independency. This will result in possibly a fourth industrial revolution allowing humans to resolve problems and not waste their time and entire lives completing business process that are repetitious.
But what about humans losing their jobs? With the already massive unemployment problem across the globe, the evolution of AI means humans at work being replaced with intelligent robots? Where is that individual supposed to earn an income and place food on the table for his/her children?
What if we pass?
Nebulous may be the best way to describe the pros and cons of investments. Add to that uncertainty and some feeling of speculation. Uncertainty is a common risk in all new business initiatives. Therefore, bad investments are nothing new in the financial world.
With everything new and strange surfacing in our world, the wisdom that prevails is comprehending the harsh reality that being left behind will be far more destructive – and even grimmer – in comparison to the benefits of taking the safe road for now.
You can read more from Paul McNeil on his blog TechMoralitics or follow him on Twitter @mcneil_lfc.
Humans need not apply!
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Australia License
Monthly Donation
Single Donation
Keep up! Subscribe to IA for just $5.
