What is Artificial Intelligence?

By Jay Shah, CEO at OpenDNA

There are a myriad of definitions for the phrase `’artificial intelligence (AI)`’. Today the phrase has been most commonly utilised, accompanying almost all technological innovations so much so that different people have different viewpoints towards AI.

Wikipedia defines it as:

---

Artificial intelligence (AI, also machine intelligence, MI) is intelligence demonstrated by machines, in contrast to the natural intelligence (NI) displayed by humans and other animals. In computer science, AI research is defined as the study of "intelligent agents": any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.

---

MIT describes it as “Making computational models of human behavior".

Wikipedia defines Machine Learning thus: “Machine learning is a field of computer science that gives computer systems the ability to "learn" (i.e. progressively improve performance on a specific task) with data, without being explicitly programmed.”

 All systems begin with the definition of some basic rules. As a crude example, consider the common spreadsheet: within a spreadsheet you can create basic formulas (rules) which can be repeated, thereby performing all our calculations much more quickly. 

 Using a non-technical analogy, think of a new born baby; the first lessons learnt are the basic dos and don’ts of survival (if thirsty: drink, if hungry: eat).

As our metaphorical baby grows, so we iterate to the second version of our fledgling AI; both become aware of their contexts and begin to derive meaningful output from a growing number of inter-dependent inputs and environments. The metaphorical parents can give more complex instructions (“if you eat your greens, you can have dessert!”) and our spreadsheet now has multiple sheets, each with complex rules which aggregate and process many rows of data, yielding an output.

Our next iteration needs to spot anomalies and significant events; we want our growing child to draw conclusions from its surroundings outside the context of the parental rule set. This is where our spreadsheet starts to let us down; we cannot write rules and formula for the infinite possible states and events which fit our definition of an anomaly.

Now we need to teach not the facts of correct and incorrect, but instead what correct and incorrect look like: “sharing is good, hitting is bad”. While we point our child towards the exemplary behavior of a carefully chosen role model, so we build a data model from a wealth of manually-classified training data. Defined as supervised learning, we instruct our computer to process unseen data and classify it through comparison to this model.

Supervised learning inherently requires a close eye on the output of the machine. As the quantity and quality of input data and context grows and fluctuates with time, adjustments to our models are achieved through re-training with more examples of right and wrong. This represents the state of the majority of artificial intelligence and machine learning products available today. “You can now use Siri to identify and rate songs” means “Apple has built a model of example data and trained Siri to recognize a new set of voice commands”.

Un-supervised learning – the teenage years - represents the next level of AI and machine learning. In this state, our machine is capable of identifying classification inaccuracies (mistakes) and – crucially – teaching itself how to correct those mistakes.

Artificial Intelligence can be applied to many industry sectors. For some, supervised learning is all that is necessary to solve business-level problems and often the complexity of the solution lies not in the methods and algorithms used, but more in the quantity of data which must be processed. For others, un-supervised learning is the solution to problems which only humans can currently resolve.

Now that you have a good grasp of the phrase “Artificial Intelligence”, it is clear that the potential of artificial intelligence is infinite. As we harness this power and couple it with machine learning, we have just created a driving force which inevitably has been responsible for paradigm shifts in almost every domain and industry in an unprecedented way. By deploying artificial intelligence within their business environment, many have reaped major rewards by reducing operational costs through automation, increasing productivity through the conductance of tasks, which are deemed monotonous in nature, with the help of machines at a much faster rate, and growing revenue through discovery of solutions to problems once considered unsolvable by human brain.

Realizing the magnificent capabilities of AI, companies are now striving to reap these benefits – automating tasks (cost benefit, speed of service), operating at all hours vs human 9-5 (ROI benefit) - and we are seeing noticeable transformations in AI powered businesses such as in finance performance tools, operations development, autonomous vehicles, cybersecurity optimization as well as in B2C businesses where AI is poised as a vital driver to keep up to speed with consumer change in expectations and potentially playing a mission critical role in creating great customer experiences.

While we appreciate the power of AI, we need to also realize that the not so secret ingredient that gives birth to AI is intelligent data. These datasets from reliable and relevant sources are pivotal to the success of AI as they are the primary source that builds the knowledge foundation for useful as well as up-to-date insights. Understanding this direct correlation between big data capabilities and artificial intelligence, companies can be seen making huge inroads in developing and executing initiatives in data and AI driven projects, which have given birth to many technological breakthroughs and has enabled the emergence of new business innovations from the development of ‘Sophie, the robot’ to self-driving cars. These advancements have been reshaped business landscape by supercharging performance and productivity as well as proven beneficial to the economy by restructuring employment.

Experts have termed AI as the mission critical technological advancements in the history of human race and is said to soon dominate the global economy.

 ____________________________________________________________________________

OpenDNA leverages Machine Learning, Deep Learning, Neural Networks, Natural Language Processing, Pattern Recognition and Visual Recognition Systems in order to build detailed psychographic profiles of users within any business environment. To learn more visit OpenDNA or contact the team for a free demo.