Why everyone should be interested in AI

Artificial Intelligence is set to become the focus of innovation, competition and wealth creation in the twenty-first century. This article identifies six big issues – both opportunities and threats – that will impact all of us today and in the future. These are:

1. AI will fundamentally change human-computer interaction

2. The fourth industrial revolution is underway

3. A new enlightenment is within reach

4. Work is about to be re-imagined

5. It's not guaranteed to be easy or fair

6. The future of the human race is at stake (seriously)

1. AI will fundamentally change human-computer interaction

In the short video below Google's CEO Sundar Pichai demonstrates the company's latest version of Google Assistant at the I/O developer conference in May 2018. This apparently simple application combines state of the art natural language processing and machine learning to produce an AI that is indistinguishable from a real human being.

The classic way to engage with your PC via screen, keyboard and mouse is about to change radically. New ways to get technology to do what we want by spoken commands or by anticipating our needs are already available in the form of digital assistants like Siri, Alexa and Cortana, and virtual and augmented reality interfaces. Before long keyboards and even smart phones will be relegated to the same category as stone tablets and quill pens. Get ready for cybernetic implants and wifi nano-machines that connect human neurons directly to the web.

The ability of computers to comb through mountains of unstructured data to identify trends and patterns, and to weed out useless information, are a boon for many industries, from medicine to internet search. The Stanford 2017 AI Index shows how question answering systems have become substantially better in the past three years and are now almost as good as humans. With web connectivity, instant translation and interactivity, online assistants will soon be a more reliable way to get an answer than another human being.

2. The fourth industrial revolution is underway

As sciences and technologies overlap, the lines between the physical, digital, and biological realms are blurring. Just as steam initiated the industrial revolution in the nineteenth century, oil and electricity powered mass production and electronics enabled information technology in the twentieth, AI is now the catalyst of growth and prosperity. This has become known as the Fourth Industrial Revolution, 4IR or Industry 4.0.

The market for AI tools and applications is exploding. By 2020, revenue from AI-based systems is expected to hit $50 billion, 8 times that in 2016. But this is small compared to the economic effects. Accenture Research predicts the impact of AI in the next 20 years to increase global efficiency by 38%, adding 1.7% to annual economic growth, which amounts to $14 trn pa by 2035. Even in the short term, Gartner estimates that by 2022 40% of all jobs will incorporate AI for efficiency and service gains.

Big business has already recognised this future. In the past few years the smart money has been moving dramatically into AI. Not because it’s fun or funky but because it makes sound business sense. From driving down costs and prices in established markets, to creating whole new industries and mashing together old ones, AI technologies are already leading the way in economic change. Many economists are now predicting AI as the prime source of competitive advantage and growth for the next 50 years.

3. A new enlightenment is within reach

In addition to the raw economics, AI offers unprecedented opportunities to enhance wealth, health and welfare for human beings through automation, efficiency and even generating solutions to big challenges such as climate change and poverty. Neural networks applied to text recognition and translation have already brought substantial results, providing instant translation of languages as complex as Mandarin, enhancing security by identifying faces the way humans do, impacting crime by huge improvements in spotting fraud, and supporting doctors to predict heart attacks. There seems to be no limit to what AI can do with application and training.

4. Work is about to be re-imagined

We’ve all seen the headlines: Robots and AI will destroy jobs. A 2017 report from PwC estimated that robots could take up to 30 per cent of UK jobs as soon as 2030 while the Bank of England has estimated that 15 million UK jobs may be at risk. But in reality the chances of mass unemployment are low - these changes are structural and will happen over a long period. In fact, with the right framework and leadership this evolution of the labour market could transform the way we spend most of our lives.

Humans are not very good at repetitive tasks because we get tired and make mistakes. By contrast machines do this well, which is why AI is powering automation for great benefit. In the future mundane tasks will be handled for us by intelligent assistants, smart homes and cities will optimise logistics and energy usage, and marketing will finally realise its potential by enabling fully automated and personalised distribution systems.

Occupations at risk, such as Machine Operators, Insurance Underwriters and Telemarketers, share a predictable pattern of repetitive activities which are possible to replicate through machine learning algorithms. At the other end of the scale, occupations with expertise gained after years of study or requiring a level of human interaction are very unlikely to be automated. These include Therapists, Teachers, Supervisors and Artists.

In the long run, as in the past two centuries of industrialisation, AI-driven automation is expected lead to massively enhanced productivity so that humans will no longer be required to do so many repetitive tasks. People will still work, but we will work better with the help of AI. The combination of human and machine will be normal in the workforce of the future. This will free us to focus on the social, spiritual and creative aspects of life, which machines are unlikely ever to replicate entirely.

5. It's not guaranteed to be easy or fair

While fears about killer robots may be entertaining, unintended and unforeseen consequences of artificial intelligence need attention today. Human civilisation is a social construct, and the artificial intelligences we create are are rooted in human values and built with imperfect data. Until AI exceeds human intelligence, it will inevitably suffer the abuses and faults inherent in our world, including prejudice based on age, gender, race and physical ability, criminal and political influence and personal or corporate wealth. And because AI has the power to amplify, there is a risk that it could make things worse not better.

In fact, we're already living with these problems. For example, Facebook's news feed has been famously abused to generate “fake news” and influence election results in the US and Europe. Worse, the company seemed to be entirely unaware and unprepared to deal with this when CEO Mark Zuckerberg intially dismissed the suggestion as “crazy”.

AI can also be adversely affected by various forms of data and technical bias. In particular, if a system is trained with biased data it will learn and replicate this, and even magnify the negative effects. Microsoft’s experimental Twitter chatbot Tay turned racist simply from the tweets it consumed and retweeted. It was shut down after a day of increasingly unpleasant rants!

The innovation and velocity of the tech sector means that dangers such as these are more likely to go unchecked than in more established disciplines such as science, medicine or law. Whilst the need for professional and ethical frameworks has been recognised these remain in principle and voluntary practice, and not yet in law.

6. The future of the human race is at stake

Charlie Brooker's award-winning series Black Mirror cleverly explores dystopian near futures. You can see it on Netflix and Channel 4. Here is a trailer:

The capacity for technology to solve more problems, answer more questions and innovate beyond the capacity of the human brain could be used for good or ill. Opinion is sharply divided on this, with heavyweight leaders such as Stephen Hawking, Bill Gates, Elon Musk and Mark Zuckerberg expressing concerns about if and when AI exceeds our ability to understand it, or if it becomes a superior intelligence. Sensationalism does not help but we should take care to not blindly follow or absolve ourselves of responsibility for our collective actions.

If history is a guide, the improvement of technology tends to elevate the condition of mankind and allow us to focus on higher order functions and an improved quality of life. But whilst it is one thing to use AI to predict and help solve problems; it is quite another to use it to deliberately control and act in ways that will make people unnecessary.

And what if the advent of an artificial general intelligence means that humans evolve into a new hybrid species? No-one knows if this will be for the better or worse, perhaps bringing unparalleled flourishing, or else ruthless suppression by a new and superior race?

Please follow me if you would like to read more of this series. Next post: When machines learned to learn.