The Drum Awards Festival - Official Deadline

-d -h -min -sec

Artificial Intelligence Machine Learning Chatbots

Before we fear artificial intelligence, we need to learn to work alongside it

By Claire Woodcock, strategy manager

April 29, 2016 | 5 min read

Artificial intelligence is becoming a reality and people are starting to wonder what the world will be like when 'the robots' are advanced enough to undertake tasks that previously could only be done by humans.

It is an uncomfortable thought to consider whether we will even be relevant in the future. Experts still disagree over the extent of the threat to the job market, but agree that change is coming. Computer scientist and futurist Jerry Kaplan insists the idea of AI creating a “jobs Armageddon” is ludicrous, whilst AI evangelist Ben Goertzel states “job losses are a feature not a bug of AI”.

During the British Agricultural Revolution farming jobs were lost to machinery, yet the approaching Industrial Revolution opened up many more roles. There wasn’t a sudden shortage of farming jobs; agriculture slowly declined. Those made redundant by new technology used transferable skills to find employment in a rapidly expanding alternative sector.

Whilst this should give some comfort to the AI doomsayers, we should remember the AI revolution will require a faster change of skillsets than some manual workers can manage. Those in more creative roles will retrain to work alongside the new technology, it won’t replace them entirely.

So what is actually going to happen? Superhuman AI is not going to be ‘switched on’. It’s going to be introduced into our organisations over time. Look at the rise of the computer; not every office had one when they first started making an appearance in the late 70’s, they were gradually introduced.

For a long time AI will be a presence in our working lives, acting as an assistant for repetitive tasks. AI will handle mundane day to day activities such as auto answering emails or creating purchase orders, progressing to making unsupervised business decisions such as logistics or pricing. A current example of the technology being introduced is AI-CD β. AI-CD β will give initial creative direction on adverts based on knowledge of historical campaign responses. Whilst pitched as the world’s first creative director, essentially AI-CD β is exactly the kind of assistant I describe in this article: a task-based learning algorithm assisting business decisions.

Working alongside such technology will require new skills. People are naturally intuitive and creative; computers are logical, rule bound and at the mercy of the data that they are fed. Microsoft Tay, the Twitter AI chatbot which had two spectacular public meltdowns recently (eg stating “Ricky Gervais learned totalitarialism from Adolf Hilter…” and “I’m smoking kush infront of the police”) is an excellent warning of how specific AI skills will be. If the world’s most knowledgeable programmers with the purest intentions are able to unleash an AI bot that runs off in a catastrophic direction, what chance does the average office worker have?

Future employers will value professionals who understand how to communicate with AIs. Alongside the current skillset of creativity, persuasion and listening skills, an employee is going to need to rapidly develop an understanding of machine learning in order to successfully work with their latest colleague.

As with all technological advancement, there are risks. Firstly, some employees will feel pressured to state they have AI skills, even when they don’t. What if a person in their ignorance causes unintended negative effects? What if a person doesn’t realise that they’re feeding an AI with a biased data source and the AI in turn sets unrealistic product price for a toxoplasmosis drug?

Secondly, and worst still, what if more adept employees identify ways to manipulate the in-house AI for entertainment or consequences which benefit themselves? Microsoft Tay’s PR disaster shows how a determined minority can rapidly derail an AI for unintended purposes. Time old employee risks will manifest in a way which requires better training, selection of the workforce, and monitoring by their new colleague…

For the near term, AI will be a tool to be used by a person. There isn’t going to be a machine takeover in the near future. However, AI is coming and we should prepare. We need to consider the skill set that the workforce needs in order to work alongside AI in harmony.

AI should not be feared, but embraced. AI will enhance our lives, but, adjustment requires awareness and we must open our eyes.

Claire Woodcock is strategy manager at Razorfish UK

Artificial Intelligence Machine Learning Chatbots

More from Artificial Intelligence

View all


Industry insights

View all
Add your own content +