AI in society: five sectors about to be rewritten by artificial intelligence and robots

04 November 2015 -

Robots that, until recently, were passive machines now possess extraordinary capabilities. We urgently need to decide what we actually want artificial intelligence to do

Helen Meese, guest blogger

We are probably some years off from the synthetic humans of the Channel 4 drama series Humans or the beguiling Ava of Ex-Machina. But will we soon reach “technological singularity” – the point at which artificial intelligence (AI) exceeds human intellectual capacity and control? Ray Kurzweil, director of engineering at Google, has predicted that the singularity might occur as soon as 2045.

Is it really possible that, in the next 35 years, an artificial intelligence could take over from humans? In which areas of life will robots play the most profound roles?

Right now artificial intelligence is largely confined to the laboratory, but the technological overflow is beginning to seep into many aspects of our lives.

The systems-in-development have so far combined one or two of the central goals of AI research: reasoning and learning; planning; language processing and communication; perception; knowledge; and the ability to manipulate objects. Some of these tasks, such as moving objects, have been relatively easy to solve from an engineering point of view; others such as language processing and perception (a relatively complex human characteristic) have proved trickier to replicate.

But there’s little doubt we’re accelerating towards meeting these goals, and a number of industries are already benefiting from them.

Manufacturing

Automated machines and industrial robots have been a common sight on production lines since the 1970s. These passive machines wait to be programmed with repetitive actions with no level of intelligence.

In the past decade, there have been significant advances in artificial intelligence techniques for the manufacturing industry. Computer-aided engineering tools and CAD/CAM techniques have significantly reduced design costs.

Machine learning is also being applied to industrial planning and decision-making processes to increase production efficiency. Predictive modelling and artificial neural networks been been employed to predict the power consumption of machines. UK universities are already developing intelligent automation techniques to enable robots to work alongside humans and, where tasks are repetitive, take over from them completely. Systems have been developed that can learn and anticipate where a human counterpart will be to avoid collisions.

In Japan there are plans to build 30 million robots to create a vast autonomous workforce that will increase the nation’s manufacturing productivity. Likewise, Google is hoping to dominate the AI applications market in the manufacturing industry with systems that will be able to adapt to changes in product lines, without human intervention.

Healthcare

Artificial intelligence will increasingly start to make decisions about how patient data is collected, processed and presented.

AI will assist physicians in patient diagnosis and provide solutions for treatments. Artificial intelligence will also enable patients with acute illnesses to live more fulfilling lives by integrating with their medical data and generating alerts and reminders when they need to take or order drugs. A US-based company called AiCure has developed a mobile AI-based system that uses facial recognition to remind patients to take their drugs and determine whether the right person has taken their medication at the right time.

Banking and finance

AI has the potential to transform the way we do business and commerce. Systems are already being developed (and some are even in operation) that can track user habits. These intelligent systems look at behavioural characteristics such as products purchased and financial data, and then personalise information on future products to meet and anticipate users’ unique and changing needs.

AI can be used to monitor financial market trends and stock and bond prices, processing and analysing massive volumes of information, providing financial advice to banks and forecasting market changes. These AI systems will be able to identify anomalies, flag up fraud attempts and even collect evidence for criminal prosecutions.

Education

The idea of a computer replacing a teacher has actually been around since the 1970s. In some respects using an artificial intelligence, which can assess any data, anywhere in the world, at any time, seems perfect for the classroom. In the US ‘Rocketship’ schools have already been introduced using a ‘blended learning’ programme – a mix of online-based classes and teacher-led learning. It is possible that, as AI develops, teachers’ roles will change, so that they become enablers rather than leading the class; letting the students explore for themselves whilst they provide guidance.

Watson – intelligence in the clouds

Will AI take human/robotic form in future? Probably not. I suspect that AI will be more ethereal, providing us with access to global information via vast cloud-based networks.

One of the best-known intelligent AI “beings” was Watson, the IBM AI computer designed to use natural language algorithms to answer questions. He [it?] won the American quiz game Jeopardy in 2011.

Today Watson has natural language, hypothesis-generation, and evidence-based learning capabilities, and has been utilised in sectors from finance to molecular biology and oil exploration; it’s even written its own cookbook. In just four years, Watson has gone from being the size of a bedroom to the size of “three stacked pizza boxes” and has increased its processing speed by 24 times. Where will it be in another four years?

For now, the idea of an AI that’s made in our own image is limited to the realms of science fiction. But within my lifetime we’ll see forms of artificial intelligence that have global reach, that we can interact and communicate with, and which will enhance our lives from cradle to grave.

Should we develop technology that exhibits intelligent behaviour? Yes, I think so. Will we see the ‘Rise of the Machines’? I am ever the optimist. I’d like to think that we’ll use this intelligence for the good of mankind, as a way to enhance society, not overrun it.

But as one of my favourite comic book creators once said :“with great power comes great responsibility.” The engineers and scientists who work in this field must ensure that they engage with society to develop technology that society actually wants.

Dr Helen Meese is head of engineering in society at the Institution of Mechanical Engineers (IMechE).