We are a long way from achieving 'pure' AI, says AutoTrader's David Hoyle

AI that can truly replace humans is still in the distant future, and automation has a long way to go

In the immediate future, businesses will remain focused on using machine learning to solve practical problems - but in the longer term will turn to artificial intelligence as machines begin to replace human work, said AutoTrader UK's lead data scientist and head of data science David Hoyle, in an interview with Computing (video).

Hoyle will present the opening keynote at Computing's AI & Machine Learning Live! conference next week, on the future of artificial intelligence.

The boundaries between AI and ML are getting blurred, and the terms are now almost synonymous, he told us. However, they remain distinct and separate: we are a long way from achieving AI in its purest, multi-skilled form. It is difficult to judge the timescale at which we might reach that point, but we are taking steps as research into the human brain and reinforcement learning continue.

AutoTrader itself is more focused on the machine learning side, which is "more practical," said Hoyle. It makes judgements, based on its historical data, to recommend prices for selling a car, and to work out what type of car a visitor might be looking for and what stage of the buying journey they're at.

Hoyle reiterated the importance of balanced data when it comes to training and using machine learning; even an algorithm that you are sure is neutral can be affected by external factors that introduce bias. With the spread of technology to touch every corner of modern life, this should reinforce the idea that practitioners must be extremely cautious about introducing unconscious bias.

"You have to think about the whole lifecycle of that model...not just the initial implementation," he said.

In the short term, most businesses will use machine learning as Auto Trader is, to provide a better service to customers. In the very near future, Hoyle expects that to remain focused on improving the bottom line, learning to better work out where the most profitable place in the business would be to apply the technology - rather than being "just a sexy piece of marketing."

Watch the video now for a preview of Hoyle's keynote at AI & Machine Learning Live!, and don't forget to register for the event.

What can artificial intelligence and machine learning do for you and your organisation? If you don't know yet, or want to make sure that you're not missing out, join us at Computing 's next IT Leaders' Forum: Preparing for an Automated Future. Attendance is FREE to qualifying IT leaders and senior IT pros, but places are going fast.