Artificial intelligence (AI) and data are inextricably linked. But in order to get the best results from an AI system, should you be cleaning your data?

This was a pivotal question at ZDNet’s Next Big Thing event in Sydney last week.

CNET editorial director Jason Hiner discussed the extent of data cleansing across industries.

“AI works really well when there are really clear parameters and where there is not much ambiguity and you don’t need context,” he said.

“If you give AI a lot of ambiguous data it doesn’t always know what to do with it.

“One of the ‘dirty little secrets’ of AI is that a lot of companies that are working on AI the most are hiring armies of human beings to cleanse the data, to curate the data and to crunch the data before you feed it into AI.”

He described these data cleansing roles as the potential blue-collar roles of the digital age.

By cleansing datasets, AI systems are being given clear and objective data to achieve the intended outcome.

But what happens when this isn’t the case?

“AI does not have subjectivity within it right now,” said Tech Research Asia founder and director Tim Dillon.

And this is something that will impact the immediate success of AI, said Dillon.

“Some of those subjective characteristics are still incredibly important for decision making in business.”

And while details like these still need to be ironed out, there was an overall feeling of optimism for AI during the event.

Microsoft Australia national technology officer Lee Hickin highlighted a potential roadmap for successful AI technology in Australia.

“In order to be successful, without a doubt, where we’ve seen the most success – whether it be AI, machine learning, or IoT – is when you have strong leadership that enables that ability for the organisation to understand, adapt, and take advantage of the technology,” he said.