From Blade Runner and Space Odyssey to The Matrix and I, Robot, Hollywood has been exploring the issue of how AI will change our lives for decades. But industry experts at Teradata Analytics Universe believe we're closer to an answer, as Dan Robinson reports

Artificial intelligence

How will AI change our lives? It’s hardly a new question but it feels like we’re moving closer to having the answers.

As machine learning begins to take hold in every day society – see personalisation from the likes of Facebook and Netflix – and researchers predict jobs being lost to machines, a gloomy picture is emerging.

But AI – or artificial intelligence – has its benefits too and could play a key role in Industry 4.0, the moniker for the fourth industrial revolution that describes the current trend of automation and data exchange in manufacturing technologies.

PWC research says AI will contribute $15.7bn to the global economy by 2030, and in North America particularly, it’s expected to boost GDP by 14% by that year through making products and services better.

Teradata Analytics Universe 2018, the data and analytics conference hosted by US industry giant Teradata in Las Vegas this week, featured a series of experts trying to provide some answers to a debate that shows no sign of slowing up.

 

How will AI change our lives? What is AI?

“AI is a big umbrella that means nothing and everything at the same time,” says Teradata’s chief technology officer Stephen Brobst, a former tech advisor to ex-President Barack Obama.

He prefers the phrase “deep learning”, a type of machine learning and a very small subset of the broader concept of AI.

How will AI change our lives
Teradata CTO Stephen Brobst speaks at Teradata Analytics Universe 2018 in Las Vegas

Deep learning uses multi-layered artificial neural networks to deliver state-of-the-art accuracy todetect objects, recognise speech and translate languages.

Despite the AI industry being worth $9bn in annual investment globally, Sol Rashidi, former chief data and cognitive officer at Royal Caribbean Cruise Lines, is keen to stress the technology is no “silver bullet”.

She believes the definition of AI will vary between different people within an organisation.

She says: “At the end of the day, your customer’s version of AI could be something as simple as a level of predictive modelling they think is forecasting, so they can predict things two steps from now.”

 

What is AI? Augmented intelligence

Rather than the common AI abbreviation of “artificial intelligence”, Ms Rashidi suggests we’re in a stage of “augmented intelligence”, in which we place a humanised focus on information, technology and communication practices.

“While we want convenience with everything we do, our goal is really to make things personalised and meaningful at every engagement – and this takes emotional IQ,” she adds.

How will AI change our lives
Sol Rashidi, former chief data and cognitive officer at Royal Caribbean Cruise Lines, speaks at Teradata Analytics Universe 2018 in Las Vegas

“So I think there still exists a beautiful blend between humanities and technology, and the two have co-existed in tandem to propel us into a whole new world of creativity, design and applications.”

 

 

How will AI change our lives? Examples of artificial intelligence

Driverless cars are one of the most interesting applications of AI, according to Mr Brobst, with everyone from BMW, Daimler AG and Volvo Cars to Google and Apple looking into the idea of driverless vehicles.

“The decisions being made in those autonomous cars aren’t strategic decisions,” he argues.

“They’re operational decisions. They’re asking ‘should I apply the brake right now to stop myself driving into the car in front of me?’ It’s not a strategic decision but it’s pretty damn important.

driverless cars

“A connected car is going to get a lot of data from when it crashed and didn’t crash so it’s going to get a lot smarter and better at making those decisions.

“I’ve been in a car with my young nephew driver and I’m really looking forward to autonomous vehicles because humans are bad drivers.

“I was the same as him. We all make the same mistakes as our parents made and their parents made, and so on. But cars with AI don’t do that.”

 

 

How will AI change our lives? Impact of AI on jobs

Seemingly every month a new report comes out highlighting an industry in which human jobs will be replaced by machines at some point in the future.

But for Mr Brobst, it’s more about “refining jobs” rather than taking them away.

He points to the aftermath of the industrial revolution, when previously dangerous jobs were replaced by safer roles and humans continued to play a big role in the industrial workforce despite key assistance by machinery.

“This really is a revolution taking place,” he says. “It’s as important as the agricultural revolution and the industrial revolution, which had the same impact that deep learning will have.

“There’s always a bit of a pain to get to better, though.”

Those unwilling to meet these changes head-on will be made to suffer, Mr Brobst admits.

American AI expert Jerry Kaplan suggested “AI doesn’t put people out of business, it puts skills out of business.”

And a report by Harvard Business Review predicted that, over the next decade, “AI won’t replace managers but managers who use AI will replace those who don’t”.

Mr Brobst adds: “Leverage the technology to make your operations better and that will protect jobs.

“But the value of deep learning doesn’t really happen until you put it into production. If you don’t have a path into production then it’s just a prototype and all you do is talk about it to someone.”

 

How will AI change our lives? Why is everyone talking about AI?

If a consumer was asked what Netflix, Amazon and Uber are good at, it’s unlikely AI will be near the top of the list.

Yet all those tech giants – and many of their contemporaries – have been both early and strong adopters of machine learning.

Ms Rashidi says: “Have they gone out to the market saying ‘we’re awesome AI companies?’ No, it’s just a mode of operation for them.

“It’s just a bunch of analytics and data models. They’re not saying they’re at the cutting edge of AI but other companies are doing that, which I find astonishing.”

So why all the sudden rage about AI?

AI technology, fun artificial Intelligence applications

Ms Rashidi admits there’s a “sexiness” to the concept that has encapsulated the mainstream media and grabbed wider attention.

But AI is also just the latest show in town in that regard.

She adds: “From 2004 to 2006, MDM (mobile device management) was the big thing. Then it was big data and about four years ago, it was Internet of Things (IoT).

“Now it’s AI. There always has to be a movement in this space, which is cool because it gets the attention it deserves.”

 

How will AI change our lives? Ethics of AI

The whole point of AI is to build bias into models but this has come back to bite a couple of major companies accused of creating “racist AI”.

Automatic tags in Google Photos may have proven useful in identifying cars, views from airplanes and memorable events such as graduations, but Google came under fire for tagging two black people as gorillas.

How will AI change our lives
Presentation by Brandon Purcell at Teradata Analytics Universe 2018 in Las Vegas

Brandon Purcell, principal analyst at Forrester Research, says the mistake happened because the tech giant hadn’t taken into account training data to reflect the people who use its cloud software.

When Amazon Prime’s same-day delivery service was launched in Washington DC and Atlanta in 2015, it was rolled out according to postal codes that Amazon identified as being its core customer base.

The problem was it excluded neighbourhoods with predominantly black residents, meaning 96% of white people in Atlanta had access to the service but only 41% of black people did.

Amazon later rectified the problem by rolling out across the city.

Mr Purcell says: “There’s a risk of reputational erosion, regulatory fines and loss of revenue from customers who only want to buy from companies with ethics and values that align with their own.”

He says the reasons these problems occur is because there are three types of bias.

While useful bias based on behaviour and differences in attitude between customers is helpful, human bias – in which historical prejudice is captured in data, like in the case of Amazon Prime – and algorithmic bias based on unrepresentative training data are harmful.

Mr Purcell adds: “In the case of training data, the problem occurs if we don’t give the machine enough data to learn from.

“Computers are very much like children. If you try to protect your children from seeing offensive content but they do hear it, they will try to mimic it.”