Will artificial intelligence take your job and how soon?
Maybe, but it could take 30 years, according to the AI Forum's report Artificial Intelligence: Shaping a Future New Zealand. Or it could be a lot sooner, depending on the job, says Victor Yuen from New Zealand AI developer FaceMe.
What's almost certain is, your job and the way you interact with the world will be changed by AI.
It's already happening, if you consider what happens when you google a question, or click on the next book or programme recommendation.
The AI Forum speculates there will be new jobs creating and maintaining AI technology, and new roles for people working alongside the machines — sort of like the person who comes over when the self-service machine at the supermarket is yelling at you.
It will take time for businesses to adapt; the report concludes significant labour reduction effects will be spread across decades.
Routine and easily repeatable jobs will be first to go, as they have over the past 30 years as computers invaded the workplace, but that can free people up to be more creative and productive.
Yuen isn't so sure. He thinks if one company cracks automated trucking, its competitors will have to fall into line quickly or go under.
And while past industrial revolutions increased jobs over time "there were world wars which were a big part of the job creation. I'm pretty sure we don't want to do that again."
The transition can be hard. A truck driver put out of work by a robot may not be able to retrain as a software engineer or data scientist.
FaceMe, which develops artificially intelligent digital assistants for customer service roles — such as the avatar screen at Auckland Airport built for the Ministry for Primary Industries to answer people's questions as they trek between customs and biosecurity — is taking the ethical issues seriously.
Yuen is a New Zealand representative on the Partnership for AI, which brings together start-ups and major companies like Facebook, Open AI, IBM, and Google, and attended the April launch in New York of its Working Group on AI, Labour and the Economy.
He endorses the partnership's vision that doing AI right requires collaboration across borders and a commitment to sharing the benefits.
Yuen says FaceMe doesn't want its digital humans to do real humans out of a job, or work with customers who want to shut down their call centres.
"The way we want to work is with people, so an AI digital human is not designed to do a human's job, but works alongside a human.
"The drivers from our customers are things like, if you want to talk to a business banker at 10 o'clock at night, it's unlikely a bank will offer that service. But there are challenger banks out there doing self-service, high-end customer experience start-up stuff and those large corporations that want to survive will have to use our technology to provide that service at night.
"If you have a digital employee delivering customer service at 10pm, would that job have been created for a person?"
He says a person working alongside AI can deliver much better quality of service and "you have the empathy, you have the ability to adapt and advance the conversation in ways which AI can't do at the moment."
At the moment there is no single universally agreed definition for AI, which is reflected in the wide range of companies the AI Forum considers part of the sector in this country.
Its definition of what's labelled "narrow AI" is "advanced digital technologies that enable machines to reproduce or surpass abilities that would require intelligence if humans were to perform them."
That could be performing specific tasks like engaging with customers, analysing MRI scans, recommending the next movie or playing chess.
The theoretical scenario where a machine could successfully perform any intellectual task as well as, or better than, a human is called "general AI", and experts say that's still decades away.
The report takes an upbeat view of what AI adoption can do in the short to medium term:
● improve productivity or efficiency by completing routine tasks faster and with more consistency than a human can;
● make sense out of huge amounts of data that human brains aren't equipped to analyse;
● improve decision-making by reducing bias, taking proper account and weighing all the facts;
● improve customer experience, with AI driven conversational interfaces known as chatbots providing faster, more accurate customer service in many languages;
● augment human intelligence, for example by curating content, using contextual text analysis and extracting the meaning from data using machine learning.
A survey of potential New Zealand users found the main driver for AI adoption for 80 per cent of respondents was to make sense of the data generated in their business, three quarters thought it could automate tedious or dangerous work, and a similar amount thought automation could reduce costs.
Some 68 per cent of respondents were considering AI to improve customer relationship management, 61 per cent plan to deploy chatbots, and 62 per cent were looking for AI products that supported financial analytics and management reporting.
Most of the talk about AI is among specialised technology staff — only a third of firms reported discussions at board level.
Yuen says the nature of AI is that technology providers like FaceMe (and Facebook) will have access to huge amounts of data about their customers' customers.
"So how do we bake into our products the ideas of transparency and privacy?" he asks.