Machines, not people, will make business decisions, says GfK head of data and science

'Not feasible' for humans to make split-second marketing decisions like a machine can, GfK's Norbert Wirth tells Computing

Artificially intelligent machines will make the majority of marketing decisions in future, because machine learning will enable them to decide which adverts to offer people in seconds.

That's what Norbert Wirth, global head of data and science at global marketing and research firm GfK, told Computing at the Yandex Data Factory Machine Learning and Big Data: Business Challenges conference in Berlin.

Speaking at the same event, Amazon director of machine learning Ralf Herbrich described how machine learning already improves Amazon's services.

While humans will continue to have a "very important" role in making marketing decisions, it would "be absolutely not feasible" for them to make them as quickly and efficiently as machine learning algorithms could, Wirth argued, describing a test he previously ran.

"I did a very interesting exercise at a conference where I was talking to marketers and researchers. I asked them to look into the future and what did they think; how many marketing decisions will be taken by humans, and how many marketing decisions will be taken by machines? People were leaning heavily towards the majority being taken by humans," he explained.

However, Wirth believes that it will be machines, rather than humans, that will be taking the important decisions in the field of marketing.

"When it comes to developing the mechanics, developing the rules, yeah, it's humans that are important," he said.

"But when it comes to applying those and taking these decisions which happen in a split second, they're not taken by humans, that'd be ridiculous; it would be absolutely not feasible. That's for machines," Wirth added.

After all, when you see a banner advert on a website, it hasn't been selected to appeal to you by a human, but by a machine, which has chosen it on the basis of your interests and previous viewing patterns.

"The online algorithm which decides whether you see a certain banner advert on a website or not, it's not a human thinking ‘now this user is going to that website, shall we show them a banner ad for Porsche or show them a banner ad for Coke?'," Wirth said. "That would be ridiculous".

Wirth argued that there will still be a place for humans, but they will be inputting the algorithims that the machine uses to ultimately make decisions.

"It's the human that says for a Coke, we'll offer that profile and we'll pay this, and for Porsche we want these profiles and we'll pay that. But with this knowledge, the algorithm ultimately takes the decision," said Wirth.

Gartner analyst Stephen Prentice has previously suggested artificial intelligence could make a better workplace manager than a human because "it always makes the right decision and never has a bad day" and therefore has the potential to improve productivity and end workplace discrimination.