By 2020, Forrester predict insights-driven businesses will steal $1.2 trillion every year from their less-informed peers. It’s no surprise that many IT leaders are being asked, “What’s our AI strategy”? BBC2’s tech visionary, Jamie Bartlett, draws up the big picture view of AI’s economic impact. Our own expert, Jason Normanton, then outlines how AI & Machine Learning in the Cloud are already helping different sectors carry out super-charged data analysis, quickly, and at scale.
Figure 4: The Impact of AI on profits by industry, Accenture
AI & the potential ‘jobs apocalypse’, Jamie Bartlett
Few subjects have leapt into the public imagination as quickly as artificial intelligence. The prospect of a sudden acceleration in the capability of various forms of A.I. is sparking a major debate about whether this could lead to a world without work.
According to the Bank of England, as many as 15 million British jobs might disappear from the twin forces of AI and automation within a generation.
Although science fiction movies focus on sentient robots, the A.I. revolution comes mainly in the form of machine learning algorithms. This essentially means giving a machine lots of examples from which it can learn how to mimic human behaviour (hence artificial).
It relies on data to improve, which creates a powerful feedback loop: more data fed in makes it smarter, which allows it to make more sense of new data, which makes it smarter, and on and on.
We’re entering exactly this feedback loop. Over the last year or so various form of machine learning has proved itself more than equal to humans at: brick-laying, fruit-picking, burger-flipping. It’s making fast inroads into more technical and skilled work, like lorry driving, and even Poker. Humans will never win at Go or Chess again.
A.I. can already outperform the best doctors at diagnosing illness from CT scans, by running through millions of correct and thousands of incorrect examples real life doctors have produced over the years.
Our Data Fast Start Workshops can help organisations of all sizes quickly get to grips with encryption, compliance, performance and data tiering to understand how AI can help gain valuable insight from your Big Data
How will AI affect the economy?
David Autor, an MIT economist, predicts society will head toward a ‘bar-belled shaped economy’. A few well-paid lucrative tech jobs at the top of the market, but many of the middling jobs – trucking, manufacturing – withering away and replaced by the jobs that can’t be automated: the low paid service sector, and millions competing for them. The pain and humiliation felt by one’s career vanishing. Are those 3 million truckers or burger flippers or fruit pickers going to retrain overnight?
And then what of the possibility of growing inequality, between the tech wizards, the innovators, who own the assets and the tech, and the rest of us? What that does for people’s sense of fairness, or mobility, or belief in society?
In the end, these will be questions for politics to answer, as well as technologists.
AI technologies CIOs need to start thinking about, Jason Normanton
Gartner argues some vendors are over-egging the artificial intelligence capabilities of their products to cash in on the “gold rush” around the technology. As a result, it can be hard to navigate a clear path through the mass of AI options available.
6 key AI technologies
OCSL identifies the following six key AI technology areas which are being utilised within Enterprises now:
- Machine Learning, Deep Learning and Neural Networks
- Natural Language Processing, speech recognition and text to speech
- Computer Vision
- Machine Reasoning, Decision making and algorithmic processing
- Targeted Business Analytics
- Robots and Sensors, IOT Applications
Machine learning, deep learning and neural networks can be used to “solve business problems through the extraction of knowledge from data”.
Deep learning and neural networks are both evolved from machine learning. Deep learning expands machine learning by discovering intermediate representations of data points and applying them to issues to potentially solve complex problems.
For example, deep learning can be used to assess a company’s supply chain or customer interactions. It can be used to identify “choke points” within processes or highlight an issue with an individual supplier.
What is a DNN neural network?
Feedforward deep neural network (DNN) is the most common type of deep learning and uses multiple layers of interconnected processing units to “discover” trends within input data.
Training a DNN, which may process thousands or millions of individual data types, relies on a highly iterative and computationally intensive process using “gradient descent” and backpropagation”. These are heuristic, numerical optimisation techniques often using large grids of CPUs to provide the computational power required.
An example of a DNN is IBM Watson. This is an AI platform for health “trained” in the health service to take “input data” in the form of symptoms, heart rate, blood pressures and laboratory results to spot “trends” otherwise known as “medical conditions”.
Some of the leading commercial application areas for AI are;
- Bots, Chatbots and Virtual Assistants embedded within websites and applications
- Conversational AI platforms
- Analytics and predictive analytics models
- Smart Objects, Sensors and Environmental controls
It’s important to start thinking about your AI strategy today, to drive future competitive advantage, tomorrow.
To take advantage of emerging technologies, such as AI, it’s vital to have the right infrastructure in place, aligned to the right strategy. A Data Assessment can be the first step to defining your future AI strategy.