Will AI change contextual analysis as we know it?
In 2016, history was made. Lee Sedol, heralded as one of the most accomplished professional Go players of recent decades, was defeated by a rookie in front of 80 million spectators. While Sedol slumped in his chair, defeated and resigned, his opponent remained silent. How had experts, reporters, and even Sedol himself been so wrong in their predictions leading up to the match?
The future hides behind a complex shroud of uncertainty, reconfiguring itself continuously like a Rubik’s Cube on an infinite loop. The farther we look into the future, the less we know. Through contextual analysis we can build an understanding of what is pulling us towards the future. As a tool, contextual analysis doesn’t solve the Rubik’s Cube but it allows us to predict and understand the patterns and mechanisms which cause the Rubik’s Cube to change.
In the 1980s and 90s, trend gurus Faith Popcorn and John Naisbitt introduced concepts such as “Cocooning” and “High Tech, High Touch” to describe emerging societal behaviors. Their insights were built on gritty and time-consuming research: countless hours spent sifting through piles of magazines and newspapers.
Visible events, the kinds of news stories that meet us every day, create a chaotic surface, giving no sense of direction and hiding most patterns. What Popcorn and Naisbitt understood was that this chaotic surface belies deeper, more slow-moving trends with the potential to transform society. By piecing together many small clues from lifestyle magazines and local newspapers, they were able to build a larger picture of societal change. The work was tedious and slow, but the method of spotting trends bottom-up has inspired more recent, data-driven approaches.
In the early 2000s a new digital era began, opening the floodgate for human potential. Vast amounts of data were created, growing exponentially with every click of the mouse. In 2010, the total information volume passed one zettabyte, roughly equivalent to one hundred million Library of Congress. However, this information remained unstructured and unavailable to most, and our ability to use it for analysis and prediction lagged behind.
Today, we are entering a new age, riding on the wave of advancements in Artificial Intelligence and Machine Learning. Through the combination of improved algorithms, access to enormous data volumes in digital form, and scalable processing power in the cloud, our ability to analyze and predict has drastically transformed.
What does this entail for contextual analysis?
Artificial Intelligence and “knowledge graphs” have drastically improved our ability to search and analyze, not only at the chaotic level of individual events, but at the level of underlying societal currents. Natural language processing, image recognition, and speech-to-text translation have expanded the domain in which data-driven trendspotting can be carried out. Self-improving algorithms enhance our ability to separate signal from noise with each try.
When Lee Sedol left the densely packed room, his opponent stayed still. Silent. As reporters and experts hurried out of the room to discuss and analyze what had just taken place, a member of Google’s DeepMind team disconnected the victorious AlphaGo, closed the screen, and gently carried it out the room. Artificial Intelligence had won the day.
By Gustaf Jungnelius and Tomas Larsson
Interested in knowing more about AI-based contextual analysis? Dcipher is an unique Agile Analytics-platform built on Kairos Future's patented algorithms. Dcipher helps you automatically find the most important trends and patterns in any text, regardless of source or data length. Dcipher can work with texts in any language, and finds, without any bias, the most relevant insights, trends, and patterns. Contact Tomas Larsson.