Australian technology entrepreneur Anthony Goldbloom has a warning for accountants, lawyers and radiologists: your jobs will be taken over by machines.

In a TED talk launched online last week and quickly viewed by 350,000 people, the founder of San Francisco-based data scientist platform Kaggle suggests a simple test if you've ever wondered if your job is vulnerable to digital disruption.

"To what extent is that job reducible to frequent high-volume tasks and to what extent does it involve tackling novel situations?" Mr Goldbloom asked in his address, titled The jobs we'll lose to machines — and the ones we won't.

There is both good and bad news for white collar professions, he told The Australian Financial Review in an interview.

Routine company audits, writing and reviewing "boiler plate" legal contracts and inspecting medical images appear vulnerable to automation.

"Given the right data, machines are going to outperform humans at tasks like this," Mr Goldbloom said.

On the other hand, tax accountants and lawyers will be needed for complex tax structuring, to take into account a company's circumstances and interpret legislation.

Similarly, creative and strategic professionals will be in demand to come up with marketing campaigns and business strategies to find gaps in the market.

Machine learning is a powerful form of artificial intelligence by which computers observe data and learn to mimic activities humans can do.

It has evolved from sorting the mail at the post office and undertaking basic credit checks for banks in the 1990s to far more complex tasks.

A review of 702 occupations by economists at England's Oxford University in 2013 found that 47 per cent of American jobs were at risk from computerisation.

Machines, however, are far less effective at tackling novel situations and not so good at tasks they have not witnessed many times before.

"Humans have the ability to connect seemingly disparate threads to solve problems we've never seen before," Mr Goldbloom said.

"This puts a fundamental limit on the human tasks that machines will automate."

A report by management consultant McKinsey concludes that the hardest activities to automate with current technologies are those that involve managing and developing people or that apply expertise to decision-making, planning or creative work.

"These activities, often characterised as knowledge work, can be as varied as coding software, creating menus, or writing promotional materials," McKinsey US-based partners Michael and James Manyikais wrote in July.

A McKinsey analysis of 2000 work activities concluded that automation has the potential to go well beyond routine manufacturing activities and transform finance and healthcare.

About half the work time of finance and insurance professionals is spent collecting and processing data, where McKinsey says the potential for automation is high.

Hence, some 43 per cent of finance activities could be automated, with mortgage brokers, financial sales agents and bank tellers particularly exposed.

Today, the big four accounting firms employ many thousands of auditors to repetitively check over the financial accounts of corporate clients.

While there are some anomaly areas that a machine may not fully understand - such as when a business tries to cheat the system - Mr Goldbloom said algorithms would be at least be smart enough to flag it for a human to check over.

"But the repetitive part is hard to imagine the big four accounting firms will hire as many auditors in 15 years time," he said.

Kaggle, a data scientist platform with thousands of community members who compete for monetary prizes in online challenges to solve problems for businesses, was founded by the former Reserve Bank of Australia economist about six years ago.

It hosted a competition in 2012 to grade high school essays. The winning algorithm matched the grades given by teachers.

Last year, Kaggle contestants were able to review eye scans and match the diagnosis of eye disease by ophthalmologists.

In this evolving labour market, the upside is that the eradication of jobs involving frequent, high-volume tasks, means the the remaining duties may be more interesting.

"There will be fewer people in these professions but the jobs will actually get better because the repetitive or tedious parts get taken away," Mr Goldbloom said.

This robot lawyer helps you fight parking tickets and avoid eviction.

Manufacturing jobs returning to the US might go to robots, not humans.

The mystifying technology paradox everyone's worried about.

You don't need to buy a Tesla to experience automated driving.

How to invest in self-driving vehicles without shorting horses.