Emea Content Editor, Computer Weekly
Human workers need to be ready to change how they work because artificial intelligence (AI) learns to do things faster, more accurately and at a lower cost – and the risk posed to organisations not adopting the technology becomes greater than the risk of being an early user.
* remove unnecessary class from ul
$(“#inlineregform”).find( “ul” ).removeClass(“default-list”);
* Replace “errorMessageInput” class with “sign-up-error-msg” class
$(this).removeClass(“errorMessageInput hidden”).addClass(“sign-up-error-msg hidden”);
* when validation function is called, replace “errorMessageInput” with “sign-up-error-msg”
* before return
function validateThis(v, form)
var validateReturn = urValidation.validate(v, form);
* DoC pop-up window js – included in moScripts.js which is not included in responsive page
window.open(this.href, “Consent”, “width=500,height=600,scrollbars=1”);
At its annual summit, IPsoft updated customers on the latest qualifications of its cognitive agent, Amelia.
AI will be used to help solve complex problems facing humanity, such as searching for a cure for cancer and building machines that can take people to Mars and beyond, but at the same time, AI is quickly learning to do a great number of more mundane human jobs.
Perhaps it is not as headline-grabbing, but the transformations in how organisations operate will be vital if humans are to push the boundaries of their existence. Whereas machines took over the activities of manual workers during the industrial revolution of the 18th and 19th centuries, AI is now doing the same for white-collar professionals – and it goes way beyond basic tasks to those done by knowledge workers.
Even medics are able to cut out large chunks of their work by using AI. For example, according to the Annals of oncology journal, scientists trained an AI using 100,000 images of melanoma cancer cells, and normal birthmarks or moles. It was then tested using 300 dermoscopic images, and 100 of the most difficult photos from this sample were handed to dermatologists to assess, so that their conclusions could be compared with the results from the AI.
The doctors, more than half of whom had over five years’ experience, correctly classified 87% of the melanomas while the AI identified 95% of the cancer samples accurately. This use of the technology frees up time for medics to focus on matters such as beating cancer, rather than just spotting it in its early stages.
So the digital worker has arrived and it is clearly more than an automated food-ordering bot. In a large enterprise, imagine a member of staff that can understand and communicate in many languages, turn his or her hand to HR requests and IT helpdesk alerts, working 24 hours a day in multiple locations. Simultaneously, the employee is undertaking an extensive programme of education just so he or she can do more for you.
This technology is available today through a platform-based AI that can be trained to be used in whatever way the user wishes. It can never be human, but can make humans superhuman.
One example is IPsoft’s cognitive agent Amelia, which can be a banker, a tech expert, an HR manager or whatever a business wants it to be.
The AI platform can now answer questions and solve problems ranging from technical issues to employees’ holiday requests described to it in natural language, whether it be verbally or through text. Imagine asking the system why the network had slowed down and being given the answer almost immediately – and an offer to fix it.
This is what Amelia offers, although it is usually the case that Amelia alerts the technician to the problem in the first place. Then, when it’s fixed, the technician can ask Amelia to look at his or her diary and find a good time to take a couple of days off. Or it could be a food-ordering interface if a fast-food chain wanted to put it at its front end.
It is the ability to understand natural language, combined with its ability to acquire and retain knowledge, that makes Amelia something like the HAL 9000 in the science fiction classic 2001: A Space Odyssey. Before going off-script, HAL was monitoring all spacecraft systems as well as astronaut wellbeing, with communication through voice, and even taking food orders.
This type of digital worker in enterprises holds the promise of cutting costs and improving productivity dramatically.
Gartner predicts that in 2018, 500 million users will save two hours a day because of AI-powered tools. And the financial savings are unprecedented when it comes to automating knowledge workers.
McKinsey estimated recently that the automation of knowledge work will enable the creation of between $5.2tn and $6.7tn globally in extra economic value. For example, bankers, wealth managers and healthcare specialists – not just personal assistants – are a few examples of the roles where chunks of work are being automated through AI.
For example, a mortgage can be generated from an application in a matter of hours rather than a few weeks using AI tools that can automatically check what needs to be checked.
All this is clearly not lost on businesses and developers. According to the McKinsey Global Institute, European organisations spent between $3bn and $4bn on AI in 2016, investment in North America reached $23bn, and China spent even more.
Chetan Dube, CEO at IPsoft and former professor at New York University, says: “Never before has a revolution of this magnitude hit the planet in just quantitative terms, not hyperbole.
“But is this hype, or is this real? Is this Westworld – is it science fiction?”
At its annual Digital Workforce Summit, IPsoft and its customers gathered to answer some of these questions by sharing their experiences of creating and using Amelia. There was a long list of customers, including Spanish bank BBVA, the UN, insurance giant AIG and mobile network operator Vodafone, and the attendance of leading scientists was evidence that today’s businesses are using technology that was previously only found alongside brains in jars in the laboratories of the world’s universities.
But it is no longer the case that you have to be highly technical or scientific to control and develop the AI. Simple tell it what you want to be done and it will try to do it and learn from the experience, even if it needs human help to complete it the first time.
The ability to understand and respond to natural language is one of Amelia’s selling points. Imagine Amazon’s chatbot Alexa or Apple’s Siri solving complex IT issues or completing important business processes, rather than turning the lights off or explaining Faraday’s law of induction and then telling you the capital of the Turks and Caicos Islands.
But starting to use the technology will not be that simple. The accelerated pace of the technology’s development means that it will be a massive leap for organisations. But IPsoft’s Dube says it is a giant leap that is less risky than holding back and waiting to see how your competitors do.
Vodafone and Amelia
A couple of years ago, mobile network operator Vodafone began using Amelia to support its IT services desk. Speaking at the IPsoft summit, Karen Brunet, technology shared services director at Vodafone, said that after early success, the company created a dedicated team to work with IPsoft to develop the technology further.
This requires a different approach to what many IT departments are used to because, unlike most software, AI does not require programming but needs to be trained, she said.
Amelia is now available in seven of Vodafone’s countries of operation, in three languages. Currently, 58% of contact to its IT services desk goes through Amelia.
“Right now, we have 20,000 chats with Amelia a month and 53% of chats are completed by Amelia without any human intervention and in the others, Amelia hands over to a human agent,” says Brunet. “We believe that, in the next few months, we are going to increase this autonomy and think we will be closer to 65% completed by Amelia.”
For the business case, the company piloted the use of Amelia on four different journeys through the service desk. Brunet says users should focus the roll-out of this type of technology on high-volume processes. “If you don’t have the volumes, you have a nice tech toy, but not the benefits to the business.”
The company achieved a return on its investment in Amelia through savings in just 24 months, she says.
All the operations where Amelia was implemented were offshored already, so the impact on internal staff was minimal.
This might not be the case for other businesses taking up the technology. IPsoft often refers to Amelia as the most human of AI platforms, but what about the humans currently occupying Amelia’s future roles?
Effects on the workforce
Speakers from the US financial industry took the stage to talk about their experiences with AI, and made no bones about its potential to replace people. Three speakers agreed that one-third of the average enterprise workforce could and should be cut immediately when AI is fully rolled out.
Although sentiment might be less severe in Europe, this is a clear direction of travel. In the IT sector, what happens in the US usually happens in the UK some time later. But the speakers said they expected a further third of workers to benefit from learning to work with AI and expanding their own horizons, while the final third of staff will have roles alongside the AI.
It is disruption of this type that means that although AI is already “an unstoppable freight train”, as one user at the IPsoft event described it, there is much for humans to learn and prepare for to smooth the coexistence of robots and people.
This is a challenge that Dube recognised: “On one side it’s curing world hunger and cancer, and on the other side it’s the dystopian final invention.”
But the promise of AI creating many new jobs is a positive counter-argument. According to Gartner, by 2020, AI will create more jobs than it eliminates.
Humans have to train Amelia to work, after all, and by freeing up humankind from mundane tasks, who knows what future roles will be needed? According to a US Department of Labor report, 65% of today’s schoolchildren will eventually do jobs that don’t yet exist, with software robots as colleagues doing the mundane work.
And coexisting with robots may soon go beyond interaction with software, if one of Dube’s predictions is realised. By 2025, he says, people will walk past an android in the corridor at work and not realise it is not human.