“Artificial intelligence” and “augmented intelligence” may be the buzzwords of today, but there is still a lot of confusion about what they are, what they do, and how they differ from – and complement – each other. Here we explore their massive potential to revolutionize research and show how GfK is leading the way in embracing the technology.
With artificial intelligence (AI), there are three options:
Supervised Learning is where the algorithm learns from training data provided by humans.
In Unsupervised Learning, the algorithm is left to discover patterns from datasets.
Re-enforcement Learning is where humans provide rules (like those of a game) to the algorithms which then “play” back and forth, learning how to succeed with those rules and data.
In all these cases, the machine’s view is limited by the data at hand as AI lacks “real world” general knowledge. The way to maximize AI is to empower people who have this knowledge.
Augmented intelligence then takes tools created by AI techniques and is used by humans to enhance the results (hence the term ‘augmented’). The role of humans is to bring more knowledge and experience to the process than is present in the dataset. The machine learning algorithms incorporate this input from humans, usually in an iterative process, resulting in an improved final classification or predictive model. These enhanced models represent the real world better and produce far greater accuracy, hence better decisions are made.
This is about a collaboration between people and machines – making products and services better but not replacing the human touch. One way to view augmented intelligence is to think of accelerating human intelligence, with machine speed and memory, to a superhuman level. Humans bring creativity and wide domain knowledge, and this is complemented by the perfect memory and computation skills of machines.
In business, augmented intelligence provides the