After decades of speculation and justifiable anxiety about the social implications for humankind, the AI (artificial intelligence) era is finally here. In the past two decades, we have seen chess grandmasters and the best jeopardy players in the world alike fall in competition to computers. The potential of Artificial Intelligence is enormous and in fact that AI could take over nearly half of all jobs in the world in the near future shows the magnanimous scope of the technology. With our lives becoming datafied, and more and more aspects of our lives and work that generate vast amounts of data, it also becomes more predictable. From SIRI to self-driving cars, artificial intelligence (AI) is progressing rapidly. While science fiction often portrays AI as robots with human-like characteristics, AI can encompass anything from Google’s search to autonomous weapons. The focuses with AI is not only on creating software that can gather knowledge, reason, learn, plan intelligently but also communicate, perceive and manipulate objects.
Not all solutions are created equal, in fact few of them really doesn’t qualify as “Artificial Intelligence” at all. For example, while analytics platforms are valuable in their own right, they cannot be classified as AI. This is largely because they are completely dependent upon human experts to make observations, test relationships between data sets, and draw conclusions. Also, predictive analytics is only good as the data it used and while it can give you useful insights, they are often compartmentalized in and of itself, an analytics platform cannot provide proactive solutions. A genuine AI on the other hand is significantly more comprehensive in both its scope and its applications. By leveraging predictive analytics AI uses machine learning to work backwards from the results, uncovering the complex issues driving a particular desired outcome. Importantly, it does so in real-time, making observations across all data inputs and adjusting its behavior with minimal need for supervision.
The general idea behind artificial intelligence is the ability for machines to think for us, whether they are projecting sales forecasts for directing a factory assembly line. However, just because a machine can “think” doesn’t mean it can do so safely. AI technology already exists in many industries, from automotive to healthcare. Machines can communicate with humans as well as each other to make decisions and follow through on processes. However, general intelligence—the ability to make complex decisions without human programming—still resides well into the future. Artificial Intelligence’s most common application is about finding patterns in enormous quantities of data. This allows companies to automate and improve complex descriptive, predictive and prescriptive analytical tasks. All the major tech companies are working on developing Artificial Intelligence solutions, ranging from Google, Facebook and Twitter and they are working hard to crack AI.
As artificial intelligence becomes more prevalent across industries, from defence to aerospace, experts in the field will focus on making AI as safe and useful as possible. Artificial intelligence will eventually touch nearly every industry on the planet, but self-driving cars are among the most sought-after developments for this technology. “Deep learning” holds the secret to unlocking AI for the automotive industry. Essentially, deep learning means that a computer delves below the surface of algorithms and coding to get smarter as it collects data. Self-driving cars would require vehicles to take into consideration numerous road factors, including:
- The location of potholes and other road hazards
- Slight curves as well as full turns
- Traffic lights, give way signs, no-overtaking zones, and other traffic signals and signs
- Traffic congestion
- Proximity to other vehicles as well as pedestrians
- Behaviors of other motorists
Simply, we can compare deep learning to the cognitive evolution of children. Kids learn as their parents and educators correct their mistakes and provide them with guidance. Similarly, computers can learn new facts and behaviors through interaction with human beings.
Many of the AI applications currently in use or development stem from the demand for automation in all industries. When companies can automate critical tasks, they reduce man hours and increase both efficiency and accuracy because of the removal of human error. Artificial Intelligence has already made promises for decades and only recently we are seeing some, significant, results. There are many challenges involved in creating truly intelligent software and machines. So in totality, for companies being able to fully extract benefit of Artificial Intelligence they would need to have a good understanding of the data generated by their own organization. AI requires vast amounts of data and such a lack of knowledge about the available data slows down the acceptance of AI applications.
Given the explosion of data from applications and Internet of Things (IoT) sensors, and the need for real-time decision making, AI is quickly becoming a key requirement and differentiator for major cloud providers. As a result, the adoption of machine learning in the enterprise may be closer than predicted as leading cloud providers are making AI more accessible “as-a-Service” via open source platforms. AI in the cloud may well be “the next great disrupter” and opens up opportunities for businesses to create powerful new AI applications fast, without building the tools, infrastructure or expertise in house. For example IBM’s Watson Developer Cloud enables developers to incorporate Watson intelligence in their apps and provides its Watson AI engine as an analytics cloud service. Consider the following complex problems in the transportation industry. Shipping companies, such as FedEx and UPS, want to figure out the most efficient and cost-effective way to deliver the most packages. Public transportation organizations need to identify city traffic patterns to keep vehicles moving without creating gridlocks. From analyzing how to fit the maximum number of packages in a delivery van, to calculating and navigating the fastest routes to deliver those packages, multiple technologies such as the IoT and big data analytics require AI to solve these complex problems.
When people think of AI, they tend to think of “human-like” or “general” intelligence. And while that may be possible in the future, today’s platforms and models are fragmented and capable of solving only very domain-specific problems. So for enterprises with various complex problems to solve, it requires multiple services from disparate platforms working together, which is why making AI technology and applications available via open sources is so critical to the enterprise. By leveraging multiple AI cloud services, companies can innovate solutions to solve an infinite number of complex problems.
Apart from the fact that AI applications are often rather disruptive technologies, which always require time before becoming mainstream. However the current rise in expectations around AI seems much more robust than it has ever been before, but it is not the AI technology alone that is feeding the frenzy. It is the environment and foundation that exist today, providing the technologies with the required data and speed. And the most exciting thing with AI is the opportunity is boundless. The data keeps growing while machines become faster so yesterday’s, today’s and the future’s AI systems can only keep succeeding.
The article was originally published on Express Computer on December 06, 2016 and is re-posted here by permission.