AI: the ultimate intern

by Richard Stirling and Hannah Miller

It doesn’t matter that AI isn’t as smart as the humans on your team. It is already capable of saving you time and energy by taking the easiest tasks off your hands, leaving you free to focus on more important things.

night-office-shirt-mail.jpg

One of the big debates of the moment is how artificial intelligence (AI) (which almost always refers to machine learning) will affect the job market. The temptation is to let one’s imagination run wild. The news is full of stories like two AIs inventing their own languages, that only 44% of adults in the OECD have skills that are better than AI, or that AI can learn in four days what it took humans 1,000 years to acquire. Hearing these, you might think that we are all about to have our jobs taken over by machines.

A fully automated future is still a long way off. Computers work best where there is a concrete decision to be made. The game ‘Go’ is a good example - it is ferociously complex to play well, but it also has a clear set of rules and parameters that decide whether a person - or programme - wins, loses, or draws the game. In most real life situations, we don’t have that kind of clarity.

Screen Shot 2017-11-03 at 11.26.10.png

People are still better than AI at making decisions based on imperfect information, or in situations with few previous examples to act as a guide to behaviour. At the height of the Cold War it took human intervention from Stanislav Petrov to avert nuclear war, by overriding the computers that falsely warned of a missile launch by the United States.

So where should we use an AI?

Despite its current limitations, AI has plenty to offer the world of work. We may be a long way from robots running our offices, but in the meantime it’s worth considering how AI could transform the way you work.

How can AI help you now? A good way of thinking about this is to imagine AI as a virtual assistant or intern, capable of the activities described at level 1 and 2 in the diagram above. AIs of this type are already making a difference across a number of sectors, including customer service, legal work, and banking, taking on data processing jobs normally tasked to junior staff and interns.

For example, Linklaters are using a proprietary AI programme called LinkRFI to classify client names, capable of processing thousands of names in a fraction of the time it would take humans. Ocado, the online supermarket, uses machine learning to optimise customer relations, building a software programme capable of triaging customer emails into those requiring no response, an automated response, or an urgent response from a human.

Like an intern, AI can take on the routine, mundane tasks that occupy lots of your time but little of your brainpower. Like the best interns, it will learn quickly from its mistakes and get better over time. And importantly, like its human counterparts, an AI intern needs clear oversight to avoid it getting carried away.

However, unlike an intern, it won’t get tired. It will merrily churn through a huge volume of tasks once it knows how to do them. You can see this in action in services like photo classification from Apple, Facebook and Google, new recommendations from Netflix, the scheduling assistant x.ai, or the wine advice I get from Naked Wines.

When you are thinking about how to harness AI in your team, think of it like an army of tireless interns and deploy where:

  • There is a defined outcome to optimise

  • There is data for the AI to learn from

  • There is a clear system in place to oversee how it works.

You don’t have to wait for the technology to advance. AI can help your team do more of what they are good at today, with a much smaller investment than you might think.

Get in touch if you want to find out how we can help your organisation make the most of AI.

AIRichard StirlingAI