There has been monumental hype about AI and how it’s going to transform the world as we know it. The issue is that AI and machine learning is such an abstract concept for most and not understanding how it effects our regular lives means it stays abstract. So what examples of AI or machine learning are you already using – even without realising it?
What Is AI?
AI at the highest level just refers to making a machine smart. It uses machine learning and deep learning to process, understand and contextualise data in order to make intelligent decisions.
AI comes in 2 main types, narrow AI and general AI. Narrow AI would be something like visual recognition of defective products on an assembly line or simple responses to customer service enquiries. General AI is something very different and instead tries to mimic a human being in the way it thinks, understands and reasons. It is capable of being trained to carry out a much larger range of tasks and learns from its experiences. Most people think Skynet from Terminator, but in reality this kind of AI doesn’t exist yet, although experts in often in fierce debate how soon it we will reach a technological singularity.
So where is AI leveraged on a day to day basis right now?
Wimbledon have a long standing relationship with IBM, who are pioneers in the AI field with “Watson”. The use of AI at Wimbledon is used for customer experience, performance enhancement and pundit analysis.
IBM have had Watson watching thousands of hours of tennis footage, reviewing stats from over 100 years of competitive tennis matches, cross referencing all the live media and social media coverage. Overall it touches millions of real-time data points, that’s a lot of data right? AI allows the consumption of all of this data and allows the machine to, in a way, understand tennis.
So what does Wimbledon do with this data?
- A cognitive chatbot allowing fans to ask questions about matches, the order of play and live scores
- Using player movement, noise level, crowd reaction and tennis understanding, Watson is able to curate and release short highlight video clips of a tennis match.
- Allow pundits and commentators to see deep insight into a match, uncovering the real reasons to a players victory.
Commuting & Route Planning
Nearly everyone who owns a smartphone will at some point rely on a mapping service to provide directions or route planning. Google Maps makes heavy use of AI to provide optimised routes for consumers. It ingests vast amounts of data from anonymised location data, providing general traffic speed and concentration, combined with user reported traffic events such as accidents or roadworks and historical journey data. Google Maps AI technology uses this data to understand what the traffic is like now as well as predicting what it is going to be like in future.
The same technology is used by companies like Uber to calculate an accurate price for a journey or minimise a wait time for pickup or even fraud detection.
Email Spam Filters
There are over 200 billion emails sent per day, and obviously there is a lot of unwanted communication within that. The first email system was born back in 1971, so you might not expect email to be leveraging an modern technology such as AI, but you’d be wrong. AI is used to start to understand what emails look like spam and which are unlikely to be important to the user. Through the use of AI and machine learning, Gmail successfully filters out 99.9% of spam.
Every applied for a loan and been turned down? Often that is down to a decision made using AI. Financial institutions must quickly decide whether to accept your application or not, understanding a large number of factors to decide if it would be an acceptable risk for the business. Considering large banks can receive hundreds of applications every minute, having a machine make intelligent decisions in a mere fraction of the time a human could is a welcome thought.
MIT researchers found that machine learning could lead to reducing a bank’s losses on bad customers by up to 25%.
Pinterest uses visual recognition and classification to allow the platform to “see” what objects are within each pin and then recommend similar pins. AI can be trained to recognise objects within an image much in the same way a human can. Humans learn what an object is by being shown it and told what is repeatedly from a young age. Over time we start to understand similarities and differences that either mean it is the same type of object or a different type. AI is taught in a similar way.
We are just scratching the surface of the uses of AI in everyday life right now, in future the implementations are beyond what we could comprehend. AI is already affecting your life in a multitude of ways, and it’s only going to become even more entwined.