Artificial Intelligence is a hot buzzword – it’s been bandied about to the point of becoming a household term, and the past few years have seen a surge in startups claiming to offer Artificial Intelligence for “X”.

According to the technology research firm Gartner, AI is 2-5 years away from mainstream adoption, and currently sits at the “peak of inflated expectations.” However, it’s a technology with truly transformative potential – Gartner also predicts that by the end of 2018 global business value derived from AI will have already reached $1.2 trillion.

Open any tech publication and AI is an all-too-common feature in the headlines, but AI also powers far less newsworthy, everyday things such as spam filters, Siri-style assistants and Netflix recommendations. Given current trends of adoption, industry leaders are positing that artificial intelligence will eventually become a part of everything we do, similar to the way that the internet has impacted everyday life.

Hype and buzzwords aside, in this post we look at 4 real-world applications of artificial intelligence today.

Wait a sec, what’s the difference between Machine Learning, Artificial Intelligence and Deep Learning?

There’s some overlap with these terms, so here’s a quick primer:

Artificial Intelligence (or AI) is a catchall term for the branch of computing that aims to simulate intelligent behaviour in computer systems. This simulation of intelligence can be as simple as the product of multiple if-then statements, or can be as advanced as a deep learning algorithm.

Machine Learning (or ML) is a subset of AI concerned with programs that modify themselves based on exposure to data. This removes the human input element typically required in programming (coming up with static rules), meaning these machine learning programs are capable of adjusting themselves based on the data they have been trained with. With this training data, these ML programs learn to minimize error or maximise a certain outcome.

Deep Learning is a Machine Learning approach that uses multiple layers of machine learning algorithms to allow algorithms to teach themselves how to classify data, and use it to judge outcomes or predictions.

1. Speech Recognition

Artificial intelligence powers the ever-popular Google Translate translation engine. No, Google don’t keep a database of every word in every language. Instead they use a sophisticated neural machine translation system, the Google Neural Machine Translation (GNMT), that works to understand the semantics of language and translates entire sentences, rather than on a word-by-word basis. This machine learning model learns and adapts over time to translate more accurately and naturally.

Artificial intelligence is also powering the rise of voice as an interface in the home environment. With Amazon’s Alexa and Google Home devices using natural language processing (NLP) to understand sophisticated user voice queries.

2. Autonomous Vehicles

While there are no road-legal fully autonomous vehicles as of yet, the partially autonomous vehicles we’ve been hearing about for the past few years would simply not be possible without recent leaps in AI, Deep Learning and sensor technology. Autonomous vehicles use deep learning – layers of machine learning algorithms – to extract instructions on how to behave from the the random and unstructured data generated from driving.

While there are many consumer pilots underway, the autonomous vehicle revolution will likely impact the commercial world first: Uber’s partially self driving trucks are already delivering freight in Arizona. These trucks are driven by software systems and sensors on the highway, while drivers take over for the last miles of the journey.

While in the port of Rotterdam, driverless trucks work tirelessly, without fatigue or the need for a break. And it’s not only the roads that will be impacted by AI: Rolls Royce is developing ‘drone ships’ – cargo ships that can be entirely controlled from land, boosting safety and efficiency at sea.

3. Medicine

AI’s ripples are being felt deeply in medicine. In August 2018 Google’s Deepmind AI system proved it could diagnosis eye disease as accurately as the world’s best ophthalmologists. While this research would require rigorous trialing and development to become an actual product, it’s another signal of AI’s huge potential for medical diagnoses, that require detecting patterns and matching them with data sets.

In another study published in npj Digital Medicine, a neural network was capable of scanning 216,000 medical health records to predict hospital readmissions, extended stays and in-hospital deaths more accurately than ever before. The biggest difference between this study and previous efforts is how the data was processed, data for these predictive studies is often cleaned up and fed into an algorithm, in this case a neural network was capable of processing unstructured raw data, and classify variables by itself to arrive at informed health outcome predictions.

4. Call Centres and Sales Enablement

There are two significant opportunities for using of artificial intelligence in call centres; 1) by providing self-service options in advance of needing to contact customer support, the number of calls that centers have to process can be reduced, while increasing customer satisfaction with quicker solutions to their issues and 2) helping call center reps to predict caller behaviour, and even using insights from AI systems to provide recommendations to call center reps on how to best resolve the customer’s issue.

In the sales realm, advances in AI are helping sales reps do what they do best: selling. Complex sales engagements require surfacing timely and relevant content to move prospects along the buying process. Rather than relying on pure gut instinct, sales enablement tools can aid this process by providing behavioural and account data that can help improve customer relationships. Artificial intelligence won’t replace the human’s role in sales and customer service anytime soon, but will remove automated, repetitive tasks and allow sellers to focus on fostering a human connection.

Conclusion

Artificial intelligence is poised to infiltrate (if it hasn’t already!) almost all the services and systems we use everyday. It’s best not to think of AI as a standalone product in itself, but as a technology that can be used in almost any setting, with initial ripple’s being felt now that will only increase in time.