"The amount of underpayment occurring now is so large that there is an effect on wages generally and on making life difficult for law-abiding employers."
Senator Sheldon said artificial intelligence could be used to detect discrepancies in payment data held by the Australian Taxation Office on employers in industries such as retail, hospitality, agriculture and construction.
"You could do it for wages and superannuation, with an algorithm used as a first flag for human intervention," he said.
The problems of underpayment are systemic and not readily resolvable just by strong law enforcement - even though that's vital.
Alistair Muir, chief executive of Sydney-based consultancy Vanteum, said it was possible to "train artificial intelligence algorithms across multiple data sets to detect wage theft as described by Senator Sheldon, without ever needing to move, un-encrypt or disclose the data itself".
Melbourne University associate professor of computing Vanessa Teague said a "simple computer program" could be designed to detect evidence of wage underpayment using the rules laid out in the award system, but that any such project should safeguard workers' privacy by requiring informed consent.
Industrial Relations Minister Christian Porter did not rule out introducing data matching as part of his wage theft crackdown and said workplace exploitation "will not be tolerated by this government".
Mr Porter said the government accepted "in principle" the recommendations of the migrant worker taskforce which included taking a "whole of government" approach and giving the Fair Work Ombudsman expanded information gathering powers.
The taskforce report said inter-governmental information sharing was "an important avenue" for identifying wage under payment and could be used to "support successful prosecutions".
In the latest case of alleged wage underpayment in the hospitality industry, the company behind the Crown casino eatery fronted by celebrity chef Heston Blumenthal, Dinner by Heston, this week applied to be wound up after failing to comply with a statutory notice requiring it to back pay staff for unpaid overtime.
It follows revelations of underpayments totalling hundreds of millions of dollars by employers including restauranteur George Calombaris' Made Establishment, Qantas, Coles, Commonwealth Bank, Bunnings, Super Retail Group and the Australian Broadcasting Corporation.
Professional services firm PwC has estimated that employers are underpaying Australian workers by $1.4 billion a year, affecting 13 per cent of the nation's workforce.
AI Group chief executive Innes Willox said the employer peak body did not "see a need" for increased governmental data collection powers.
Australian Retail Association president Russell Zimmerman said retailers were not inherently opposed to data matching as employers who paid workers correctly had "nothing to fear" but was unsure how effective or accurate the approach would be.
"We don't support wage theft," Mr Zimmerman said.
He blamed the significant underpayments self-reported in recent months on difficulties navigating the "complex" retail award.
Senator Sheldon rejected this argument, saying the system was "only complicated if you don't want to pay".
"You get paid for eight hours, then after that you get overtime and you get weekend penalty rates," he said.
Australian Council of Trade Unions assistant secretary Liam OBrien said the workplace law system was "failing workers who are suffering from systemic wage theft".
The minister, who is consulting unions and business leaders on the detail of his wage theft bill including what penalty should apply if employers fail to prevent accidental underpayment said the draft legislation should be released "early in the new year".
Dana is health and industrial relations reporter for The Sydney Morning Herald and The Age.
Originally posted here:
Fels backs calls to use artificial intelligence as wage-theft detector - The Age