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 1049

Associated Incidents

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

Loading...
Centrelink debt scandal: report reveals multiple failures in welfare system
theguardian.com · 2017

This article is more than 2 years old

This article is more than 2 years old

An ombudsman’s report on the roll out of Centrelink’s automated debt recovery service has identified multiple failures that placed unreasonable burdens on welfare recipients and staff.

The self-initiated investigation was announced in January after months of complaints that the problem-riddled system was sending incorrect debt notices to people.

“The [Online Compliance Intervention] project effectively shifted complex fact finding and data entry functions from the department to the individual and its success relied on its usability,” said the report by acting commonwealth ombudsman, Richard Glenn.

Community legal centres warn cuts will leave Centrelink’s robo-debt targets helpless Read more

In July 2016 the Department of Human Services switched to an automated system of data-matching to identify welfare fraud.

Glenn found the switch did not see an increase in the 20% rate of people who were sent debt notices but then found not to owe a debt, and it was “entirely reasonable and appropriate” for the department to ask customers to explain discrepancies.

However, this was only as long as the system had complete and accurate information, it said, and the department’s processes placed a greater emphasis on the customer’s responsibility, which was “not reasonable or fair” in many situations.

The department’s requirement for people to keep records for six or seven years was not reasonable, the report found, “particularly when they have not been forewarned about this requirement”.

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

The report also highlighted a litany of other issues related to planning, implementation, consultation, expectations of welfare recipients and staff, and a lack of understanding and communication about the new system’s complicated nature.

While it was inevitable there would be issues when rolling out “a large scale, complex automated system in a short timeframe”, the department failed to properly mitigate risks during transition or consult adequately in the planning processes.

“DHS did not clearly communicate aspects of the system to its customers and staff, which led to confusion and misunderstanding.”

The ombudsman acknowledged improvements already made by the department but said there was more needed.

The report identified key issues including:

The accuracy of debts raised, in particular those that were calculated using “averaged” income data.

The 10% recovery fee.

The transparency and usability of the OCI system.

The problems faced by customers when gathering evidence and presenting their case.

The adequacy of the department’s assistance and communication with customers.

The adequacy of staff training and communication to support customers using the system.

The department’s approach to complaints.

The adequacy of the department’s project planning and governance mechanisms.

Among its findings, the report said the department’s initial letters to welfare recipients were “unclear and deficient in many respects” and left out crucial information, including that their income would be averaged out if they didn’t enter it each fortnight.

The letters were also missing contact numbers for the compliance team, which led people to call the general line and face long wait times and unprepared staff who often did not know how the system worked.

Centrelink inquiry told 'income averaging' creating incorrect welfare debts Read more

The 21-day timeframe to respond to the initial letter was also not reasonable or fair in all circumstances.

The ombudsman’s reports concluded the implementation problems could have been mitigated with better planning and risk management, including user testing, a more incremental roll out and better communication with staff and stakeholders.

The federal government has stridently defended the debt recovery system, including going so far as to release personal details about a welfare recipient to media in order to publicly rebut complaints.

The human services minister, Alan Tudge, said the government welcomed the report and accepted all recommendations.

In a statement, Tudge repeatedly noted the parts of the report that defended the automated system and said the government was already making improvements that, in some cases, went further than what was suggested by the ombudsman.

“The unfortunate reality is that while most welfare recipients do the right thing, some deliberately defraud the system while others inadvertently fail to accurately declare their income and consequently receive an overpayment,” he said.

“We want to be fair and reasonable to welfare recipient but also fair to the taxpayer who pays for the welfare payments.”

The shadow human services minister, Linda Burney, said the report raised “serious questions about Alan Tudge’s oversight of his department”.

“While some changes have been made to Tudge’s robo-debt system, the ombudsman is clear they don’t go far enough,” sh

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