Skip to Content
logologo
AI Incident Database
Open TwitterOpen RSS FeedOpen FacebookOpen LinkedInOpen GitHub
Open Menu
Discover
Submit
  • Welcome to the AIID
  • Discover Incidents
  • Spatial View
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Submit Incident Reports
  • Submission Leaderboard
  • Blog
  • AI News Digest
  • Risk Checklists
  • Random Incident
  • Sign Up
Collapse
Discover
Submit
  • Welcome to the AIID
  • Discover Incidents
  • Spatial View
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Submit Incident Reports
  • Submission Leaderboard
  • Blog
  • AI News Digest
  • Risk Checklists
  • Random Incident
  • Sign Up
Collapse

Report 1052

Associated Incidents

Incident 5739 Report
Australian Automated Debt Assessment System Issued False Notices to Thousands

Loading...
Centrelink 'Robo-Debt' Program Unclear, Unreasonable And Unfair: Ombudsman
huffingtonpost.com.au · 2017

Centrelink failed to communicate properly with welfare recipients, their staff lacked training and its so-called 'robo-debt' program placed "unreasonable" expectations on claimants and should have been more rigorously tested.

Those are the findings of the Commonwealth Ombudsman's investigation into the Centrelink debt recovery scandal which saw thousands of people "terrified" over debt notices often totalling many thousands of dollars.

Since Centrelink's debt recovery system was automated in July last year, thousands of clients have spent many hours and days disputing debts they did not actually owe -- like this man, who spent 30 hours disputing his daughter's Centrelink debt, which was consequently amended from $5,500 to $0.

Now, the Ombudsman wants a "comprehensive evaluation" of the automated program before it is rolled out any further.

#Centrelink's debt recovery system lacks transparency and too difficult to use, ombudsman finds https://t.co/deaVKY4boK — ABC News (@abcnews) April 10, 2017

"We found there were issues with the usability and transparency of the system," Acting Commonwealth Ombudsman Richard Glenn said.

Since the investigation began in January, the Department of Human Services has made many positive changes to make the system clearer and easier to use, but "more improvements are needed to ensure the system reflects good public administration," the Ombudsman said.

But the overall accuracy of the debt assessment program was the same and the widely reported "error" rate of debt notices being sent incorrectly was a misrepresentation, the 113-page report found.

The number of times no debts were found following an initial "request for information" letter was 20 percent before the system was automated and did not increase following the automation, the Ombudsman found.

The time frame to respond to the request (21 days), on the other hand, was "not reasonable or fair" in many cases, "given the complexity of collecting historical employment information or the possibility that the customer may not have received the initial letter".

This is not reasonable or fair in situations where customers have to collect evidence from several years ago"

In February, the deadline was extended to 28 days, with options to apply for extensions, following recommendations from the investigation and complaints from welfare recipients.

The short time frame meant that people were forced to repay debt they didn't believe they owed, because they couldn't successfully dispute the debt before the three week deadline.

"The most harrowing part was being told that even though I've got only three weeks left to pay back the debt, the review takes six weeks and that I need to start paying it back in the meantime," Sydney man Brett told The Huffington Post Australia.

The federal watchdog also slammed the Department of Human Services (DHS) for putting too much onus on people to verify their earnings -- sometimes for up to six or seven years.

"This is not reasonable or fair in situations where customers have to collect evidence from several years ago, or where the customer does not have the capacity to obtain the evidence," the report reads.

"Customers do not have the same information gathering powers as DHS."

ANDREW MEARES/FAIRFAX Kathryn Campbell Secretary Department of Human Services before a Senate Committee hearing into the Centrelink robo-debt collection in Canberra.

The DHS should use its powers to collect information directly from employers to verify the customer's earnings, the report stated.

Centrelink should also consider paying back the ten per cent "debt recovery fee" which was automatically applied to debts when the automated system was introduced -- and which the Human Services Minister Alan Tudge was reportedly himself unaware of when questioned by journalists in December.

The report criticised the initial debt letters sent to customers as "unclear", blaming a failure to include the correct phone number (and the difficulty locating the phone number on the Department's website) on the long wait times Centrelink's main phone line experienced.

This resulted in 28 million calls to Centrelink getting a busy signal in the past year, with some customers calling hundreds of times before getting through.

The dedicated 1800 compliance helpline number has since been added to the initial contact letters, the report noted.

But the review has further recommended that the letters mention the possibility of debt earlier, clearly explain how to log earnings so the correct debt will be calculated, and explain to clients that they are able to apply for an extension for their debts.

The Department of Human Services has agreed to implement all of the federal watchdog's recommendations. The Ombudsman Richard Glenn said he would "continue to work closely" with the Department to monitor their implementation of the recommendations.

ALSO ON HUFFPOST AUSTRALIA

Read the Source

Research

  • Defining an “AI Incident”
  • Defining an “AI Incident Response”
  • Database Roadmap
  • Related Work
  • Download Complete Database

Project and Community

  • About
  • Contact and Follow
  • Apps and Summaries
  • Editor’s Guide

Incidents

  • All Incidents in List Form
  • Flagged Incidents
  • Submission Queue
  • Classifications View
  • Taxonomies

2024 - AI Incident Database

  • Terms of use
  • Privacy Policy
  • Open twitterOpen githubOpen rssOpen facebookOpen linkedin
  • e1b50cd