This is part 2 of our Welcome to the machine piece taken from Issue 5 of HomeWork magazine. Read part 1 here.
Until recently, machine learning was either too expensive or impossible for small and medium-sized businesses to get their hands on. These days it’s not only accessible; it’s fast becoming an imperative. “Businesses that don’t adopt it are going to be overtaken by businesses that do – it’s like an arms race,” says Doug Ayres, MD at Filament, an agency that advises and implements machine- learning techniques for clients from its office at Workspace’s Cargo Works in Waterloo.
Filament has devised conversational interfaces (chatbots) for clients including Deutsche Telekom and HSBC in America, enabling them to automate tasks in a way that simply wouldn’t have been possible a few years ago. It is currently working on a suite of tools that will allow companies to create their own machine-learning models without the need for in-house data scientists – making AI even more accessible for SMEs.
Like humans do
However, AI is no silver bullet, Ayres warns. We’re a “long way off general AI”, and so for now at least, it’s all about “computers and humans working together to do things better”. Ben Gancz, Director at Qumodo, would agree on the importance of compatibility. The company, which is based at the Print Rooms in Southwark, focuses on human interaction with AI, particularly in the defence, law-enforcement and medical sectors. One of its projects uses AI technology to help adults and children remove explicit images of themselves from the internet, saving police and moderators time. As well as developing AI tech itself, which has also been funded by Innovate UK, Qumodo is researching human psychology in order to integrate AI better into human teams. “There is maximum hype around AI at the moment,” says Gancz, “But it won’t meet people’s expectations unless there’s more investment into interfaces and getting people to have confidence in these technologies.”
At the moment we have unrealistic expectations that AI will be 100% foolproof. We are horrified by the thought that a driverless car might ever crash, while somehow okay with the idea that human errors cause crashes all the time. “People have an inflated expectation of what AI can do, and when it doesn’t do that it can really turn people off,” says Gancz. The challenge is getting people to “appropriately calibrate their trust” in AI – to have confidence in it, but not too much. “Policemen and soldiers need to really know whether what they’re being told is a good idea or not. We don’t want people to blindly trust it.”
What about the tasks that no longer require any human interaction? The loss of jobs to AI is already happening and is a major ethical issue that needs to be explored further. “Machines are going to take over a lot of repetitive tasks that people get paid for now,” says Ayres. Many fear that technology will make people redundant, depress wages and lead to a stagnation in living standards unless we take action now.
So how concerned should we be about the rise of the robots? Last year, Facebook had to shut down a pair of its AI robots after they reportedly invented their own language. Although something of an AI evangelist, Michalis Michael can’t rule out the possibility of the machines taking over. “We should be very careful not to get to a Terminator-type scenario,” he says. “Politicians should start discussing the legal and ethical frameworks around AI because at the moment there aren’t any.”
People have an inflated expectation of what AI can do, and when it doesn’t do that it can really turn people off,
Ben Gancz, Director at Qumodo
Others believe that our fears about AI and automation have been overstated. “Technology automates tasks, rather than jobs,” says Benedict Dellot, co-author of The Age of Automation, a report commissioned by the Royal Society of Arts. “If a worker loses one task to automation, it’s likely that another will come along and replace it.” Dellot was a panel speaker at Workspace’s Rise of the Bots Business Insight Dinner in February. He sees potential for AI to create jobs in the IT sector, and to complement workers, rather than replace them. For example, AI models can help doctors reach diagnosis faster, or farmers to detect and treat crop diseases.
Gancz is also optimistic. While acknowledging that machines will inevitably become superintelligent, he doesn’t think this will necessarily lead to our destruction or enslavement. He believes that in a world of automation, people will value human interaction all the more, and our desire for human experiences might see an expansion of, for example, the services industry. “People have coffee machines in their house that make very good coffee, but they still love going to coffee shops and paying five quid for someone to make it for them,” he says. “It’s the human connection that makes it valuable.”
There may be scary moments ahead, but whether you’re terrified or excited by AI, it seems unlikely to be a flash in the pan. Businesses should embrace its potential or risk being left behind.
Do you want to join our thriving community of businesses? Come and see what we have to offer your company and make your mark on the capital.
If you’re looking to rent commercial property in London, it couldn’t be simpler with Workspace. Browse our portfolio of 64 properties across London, check out our inspiring events calendar, and if you’re not ready for your own office space, you can find information on all of our flexible co-working memberships too.