How artificial intelligence has become crucial to our modern lives

Consider for a moment these questions:

• What is AI?
• How do I feel about it?

Then think how often you’ve heard or read about fears of AI – that it’s ‘taking over the world’, that it will be the ‘death of mankind’.

Sure, artificial intelligence can be a powerful tool for evil in the wrong hands – witness the recent spate of foreign interference in various democratic elections – but the fact of the matter is we’d nowadays be lost without it.

You may not be fully aware of it, but you’re almost certainly using AI every day. Some artificial intelligence examples? Product recommendations from services such as Netflix and Amazon, and virtual assistants like Siri, Cortana, Alexa and Google Assistant, just to name a few.

Data from your web experience is collected, organised, edited and then processed by the algorithms, which then continually update your likes and product preferences and send you recommendations. In the case of AI assistants, the algorithm has ingested massive amounts of speech data, then broken down, curated and learned the different phonemes and the usual responses to standard questions.

Give me an AI definition

AI is the generic term for the digital technology used to solve different types of problems, depending on the complexity, learning approach and data type. The ability to acquire and apply knowledge and skills is what we as humans associate with the word ‘intelligence’. When the same is done by computers we call it ‘artificial intelligence’.

AI primarily processes available data with the help of applied fundamental maths such as linear algebra, probability, information theory and numerical optimisation, then encapsulates it into algorithms to acquire and apply knowledge.

Here’s a breakdown of the primary classifications of artificial intelligence:                  

“Deep Learning” by Ian Goodfellow, Yoshua Bengio & Aaron Courville.

Why the rise of AI?

Although the algorithms have been around for quite some time, in the last couple of years we’re seeing an unprecedented rise in the use of AI technology. One of the main reasons for this is the sheer volume of available data being captured today. Plus the hugely increased power of computers in recent years has further augmented the power of the algorithms to use these massive amounts of data, to deliver faster and more efficient outcomes. Lastly, due to the open source revolution there’s been a massive spike in the number of advanced new algorithms, which is also increasing AI efficiency.

AI – the myths

We’ve all heard them. That AI is somehow sinister and unknowable – that it’s an unstoppable force that’s working against, not for us. Then consider the reality – the wide range of AI based products that are making our lives easier every day.

These include:

• Virtual assistants
• Video games
• Credit card fraud detection
• Loan/mortgage default prediction
• Online customer support using natural language processing
• Online purchase prediction and product recommendation
• Smart ad serving
• News generation
• Security surveillance
• Self-driven cars
• Medical imaging and diagnostics

In other words, artificial intelligence has become mainstream. Most rumours and fear of AI really do spring from ignorance and a lack of understanding. Although we harness technology to bring death and destruction to the world in a military capacity, we also use it to solve some of the great problems facing mankind.

AI is poised to power the fourth industrial revolution. It’s up to we humans how we use it – for good or for bad. The actual AI technology itself isn’t the cause of wrongdoing – it’s the people who deploy it who need the scrutiny.

What are the challenges of using AI?

Obviously, as is the case with any valuable technology, the implementation of AI does have its challenges. Here are the main ones:

• The expertise, vision and information needed aren’t always readily available
• The results delivered by AI are only as good as the data provided
• Considerable expertise and skill on the part of the data scientist is a prerequisite for success
• Each business AI project generally needs a uniquely tailored solution, depending on the data and complexity of the situation
• Regular monitoring and tuning of hyperparametres for optimal relevance and performance, each time an algorithm is trained
• A substantial span and history of quality data is necessary for a satisfactory outcome; the more data, the better.

Will AI help your business?

Your business may not need a self-driven car, or facial recognition software.

However, if:

• Your organisation is data-driven
• You’re capturing a substantial amount of data
• Processing this data has become overwhelming
• You need to build a solution to continuously learn from your data
• You want to apply these learnings to your decision making
• All of the above could unleash your productivity and increase your business value

then you should definitely consider an AI based solution. You’ll find that AI will deliver results exponentially faster, and all the data that was sitting idle can be put to good use.

For example – you’re wanting to provide extra, especially tailored service to your high value customers. Using your existing data AI will identify these customers, then harness their behaviour and spending history to customise your offerings to them.

Commissioning AI – important points to consider

Firstly, of course, you need to identify the particular business issue that you want solved. This is where our experts at Bridge can help. We can determine firstly whether an AI process will be suitable for your business case. And if the answer is ‘yes’, what type of structure you’ll need. Lastly, we can help you choose the right professionals to implement your transformative AI solution.

Have any tips or tricks on your own experiences with AI? If so, share them in the comments below.

Animesh Khan, Senior Consultant