There’s nothing artificial about our intelligence
Since the dawn of the digital computing age, mankind has been fascinated by the concept of artificial intelligence. In the twenty-first century, AI has become a part of everyday life, but perhaps not in the way the pioneers of science (and science fiction) imagined.
Whilst Charles Babbage (1791-1871) may be hailed as the father of the computer, and Ada Lovelace (1815-1852) recognized as the world’s first computer programmer, the Analytical Engine they collaborated on bears scant resemblance to what we have come to recognize as “a computer”. By contrast, the pioneering work of Alan Turing (1912-1954) in the field of computer science earned him the title “the father of artificial intelligence” and sees his name writ large throughout the modern world of computing and beyond.
More recently, Transversal cofounder Sir David MacKay made his own, significant contribution to the field. His advances in machine learning and information theory included the development of Bayesian methods for neural networks.
Science fiction’s take on AI has been dominated by the quest to build something that looks and acts “human”. It has taken the concept of artificial intelligence to the limits and interpreted it as artificial life. From Maria’s robot double in Fritz Lang’s Metropolis to the Gunslinger in Michael Crichton’s Westworld or the replicants chased down by Harrison Ford in Ridley Scott’s Blade Runner (where Turing gets a name-check), they all “believe” they are human.
One other thing these all have in common – it all ends up going badly wrong! However, the negative perception of AI in the movies doesn’t stop with things that look like people. Any machine that becomes self-aware seems intent on ending human life as we know it – SkyNet from Terminator, HAL from 2001 and the WOPR from War Games (anyone for a game of tic-tac-toe?).
Science fact is somewhat different.
From a purely scientific perspective, AI refers to devices that mimic cognitive functions. Often thought of as an ability to solve complex problems, to learn or to “reason”, AI has wide ranging implications for a big data world.
AI has subtly become a part of everyday life, with applications ranging from virtual assistants (Siri and Cortana) to self-driving cars, security surveillance to smart home devices. AI is also revolutionizing the way businesses interact with customers. From voice recognition and intelligent routeing to online customer support, the application of cognitive knowledge management is helping to improve customer self-service and streamline the customer experience.
For Transversal, AI is about processing big data in a more intuitive way. It is used to analyze user behaviour, to add context and to predict future needs, allowing tasks to be completed with as little effort as possible and creating a near “frictionless” experience.
As a tool to reduce effort, AI is an enabler for smarter data analytics. Responding to user interactions in real time, reducing the number of repetitive requests and analyzing search results or content to identify knowledge gaps much faster than a human.
Semantic search technology utilizes a bank of linguistic concepts to analyze what a user means, rather than what they type. For example, a supermarket customer searching for “stock” is more likely to want to make soup than to check the company share price. The same technology is used to power predictive searches, making recommendations for searches as a user begins to type their enquiry.
AI is also the driving force behind those related items or articles you see on a variety of websites. The more frequently a related item is clicked, the more likely it is to be returned in the future as the system learns what is truly relevant.
What AI isn’t (at least for us) is a user interface. It isn’t about chatbots, fooling customers into thinking they are interacting with a real person and forcing interactions down a narrow channel. It isn’t about mimicking speech or other human behaviour, using resources that could better be spent on knowledge management.
In short, AI is not about building a system capable of beating the Turing test (yes, him again). It’s about providing cognitive applications that support self-service or agent-assisted interactions, not replace them.