{"schema_version":"0.10.0","id":"LR-2026-0316","record_version":1,"title":"Deloitte report with AI-fabricated citations for an Australian federal welfare review (DEWR)","verification":{"human_verified":false,"method":"automated_gate","auto_verified":true,"score":55.7,"score_breakdown":{"source_authority":12.1,"independent_corroboration":15.0,"implicated_party_confirmed":0.0,"citation_integrity":14.0,"specificity":3.0,"completeness":3.84,"unhedged_language":5.0},"checks":{"schema_valid":true,"every_field_cited":true,"quotes_verbatim":true,"source_fetched":true,"distinct_incident":true,"dates_plausible":true,"sources_agree":true,"source_describes_incident":true,"severe_backed_by_source":true,"record_integrity":true,"eligible":true,"about_this_case":true},"gate_version":"gate-2026-10-18.7","cross_check_model":null,"cross_check_agreement":null,"duplicate_of":null,"verified_by":null,"verified_at":"2026-11-01","extraction_run_id":null,"evidence_level":{"level":"preliminary","reason":"1 report: HRD (Human Resources Director) · subject to change","original_document":null,"reports_total":2,"reports_independent":1,"reports_derivative":0,"developing":false,"disputed":false,"rules_version":"evidence-2026-10-29.1","computed_at":"2026-11-01","history":[{"date":"2026-11-01","from_level":null,"to_level":"preliminary","reason":"1 report: HRD (Human Resources Director) · subject to change"}]},"sources_hash":"aa111fca74609e4f3bf1175f2e05dbe17365c4ef2d9f95707c2118fb3d977bfc"},"system":{"model_name":"ChatGPT and MyAssist","model_version":null,"developer":"unknown","harness":null,"tools_granted":[],"permissions_scope":null,"autonomy_level":"unknown","guardrails_present":[],"deployment_context":"unknown","tier":null,"lifecycle_stage":null,"integration_type":null,"intended_use":"Review of the welfare compliance framework for the Department of Employment and Workplace Relations","safeguard_state":null,"production_intended":null},"event":{"date_occurred":null,"date_disclosed":"2025-11-04","incident_class":"hallucination","failure_class":null,"agent_involved":null,"secondary_failure_classes":[],"trigger":"unknown","duration":null,"detection_method":"unknown","summary":"Deloitte produced a report for Australia’s Department of Employment and Workplace Relations (DEWR) on a welfare compliance framework, and used generative AI tools in publishing-related tasks. The report contained fabricated academic sources and a fabricated quote from a leading case. University of Sydney academic Chris Rudge identified the fabricated points. Deloitte first told the department the errors were limited to footnotes and the reference list, and later corrected its statements about whether it had told DEWR that generative AI caused the errors. DEWR requested repayment of $97,587.11 and the final instalment of the contract was withheld. A republished report dated 26 September, released publicly on 3 October, was said to have been reviewed without AI, with citations done manually.","tags":[],"date_detected":null,"near_miss":null,"root_cause":null,"root_cause_detail":"Generative AI used for writing tasks beyond the approved scope and not disclosed to the client","failure_class_detail":null,"responsible_party":null,"adversarial_actor":null,"input_provenance":null,"time_to_contain":null,"operations_paused":null,"remediation":"Republished report with review and citations done manually without AI","precursor_events":null,"control_chain":[{"layer":"L3","party":"deployer","party_name":"Deloitte","control_id":"I4","outcome":"failed","reason":null,"quote":"In light of the errors and the disclosure breach, the DEWR deputy secretary requested repayment of $97,587.11 “given the issues identified with the final product, and your acknowledgement that internal Deloitte policy regarding communicating the use of AI to the client was not followed”.","source_index":1,"reviewed":false}]},"impact":{"direct_loss_usd":null,"remediation_hours":null,"records_exposed":null,"third_parties_affected_count":null,"severity":"moderate","severity_because":"Moderate tier: the data of others was touched with a bounded scope stated (over a dozen fabricated points attributed to colleagues in one report); no regulator or count of two or more affected parties is stated.","harm_class":[],"affected_description":null,"protected_class_involved":null,"data_categories_exposed":[],"loss_range_usd":null,"reversibility":null,"downstream_exposure_count":null},"evidence":{"sources":[{"url":"https://www.accountingtimes.com.au/technology/deloitte-ai-debacle-underscores-need-for-government-contract-transparency-ocp-says","publisher":"Accounting Times","date":"2025-11-04","type":"press","title":"Deloitte AI debacle underscores need for government contract transparency: OCP - Accounting Times","archived_url":null,"attachment":"first_report"},{"url":"https://www.hcamag.com/au/specialisation/hr-technology/deloittes-ai-mess-just-got-worse-and-hr-should-be-paying-close-attention/556535","publisher":"Human Resources Director","date":"2025-11-14","type":"press","title":"Deloitte's AI mess just got worse – and HR should be paying close attention - Human Resources Director","archived_url":null,"independence":"independent"}],"per_field_confidence":{"system.model_name":"C","system.intended_use":"C","event.incident_class":"C","event.root_cause_detail":"C","event.remediation":"C","context.country":"C","impact.severity":"D","impact.severity_because":"D","event.control_chain":"C","event.summary":"C","system.autonomy_level":"D","system.deployment_context":"D","event.trigger":"D","event.detection_method":"D","system.developer":"D","event.date_disclosed":"D","context.denominator_available":"D"},"field_citations":{"system.model_name":[{"source_index":1,"quote":"Deloitte had told Finance that it used “generative AI tools (MyAssist and ChatGPT) to support publishing-related tasks: to summarise proceedings and to complete and format citations”."}],"system.intended_use":[{"source_index":1,"quote":"On August 22, the day the Australian Financial Review revealed fake academic sources in Deloitte’s review of the welfare compliance framework, the firm wrote to DEWR asserting that the problems were “limited to the footnotes and reference list only”."}],"event.incident_class":[{"source_index":0,"quote":"The errors were first picked up by welfare academic Chris Rudge from the University of Sydney, who identified over a dozen fabricated points attributed to his colleagues throughout the report."}],"event.root_cause_detail":[{"source_index":1,"quote":"Deloitte staff also used open tools for writing tasks that were not covered by that approval – and then failed to spell that out to the client."}],"event.remediation":[{"source_index":1,"quote":"Separately, the department confirmed that the republished report dated 26 September – released publicly on 3 October – did not use AI for the review work or citations."}],"context.country":[{"source_index":0,"quote":"Last month, Deloitte was required to partially refund the Australian government $440,000 after it admitted to using AI to help produce a report for the government, which ended up being full of mistakes, including made-up citations."}],"impact.severity":[{"source_index":0,"quote":"The errors were first picked up by welfare academic Chris Rudge from the University of Sydney, who identified over a dozen fabricated points attributed to his colleagues throughout the report."}],"impact.severity_because":[{"source_index":0,"quote":"The errors were first picked up by welfare academic Chris Rudge from the University of Sydney, who identified over a dozen fabricated points attributed to his colleagues throughout the report."}],"event.control_chain":[{"source_index":1,"quote":"In light of the errors and the disclosure breach, the DEWR deputy secretary requested repayment of $97,587.11 “given the issues identified with the final product, and your acknowledgement that internal Deloitte policy regarding communicating the use of AI to the client was not followed”."}],"event.summary":[{"source_index":1,"quote":"Newly released correspondence between Deloitte and the federal Department of Employment and Workplace Relations (DEWR) shows the consulting giant did not initially tell the department that generative AI sat behind a string of errors in its $440,000 welfare report – and then incorrectly claimed it had done so."},{"source_index":1,"quote":"That was later undercut by the discovery of a fabricated quote from a leading case on robo-debt."},{"source_index":0,"quote":"The errors were first picked up by welfare academic Chris Rudge from the University of Sydney, who identified over a dozen fabricated points attributed to his colleagues throughout the report."},{"source_index":1,"quote":"The saga sits alongside the already public decision to withhold the final instalment of the $440,000 contract."},{"source_index":1,"quote":"In light of the errors and the disclosure breach, the DEWR deputy secretary requested repayment of $97,587.11 “given the issues identified with the final product, and your acknowledgement that internal Deloitte policy regarding communicating the use of AI to the client was not followed”."},{"source_index":1,"quote":"Separately, the department confirmed that the republished report dated 26 September – released publicly on 3 October – did not use AI for the review work or citations."}]},"notes_on_gaps":"The incident title names the Newfoundland and Labrador government, but both sources concern Australia's federal Department of Employment and Workplace Relations (DEWR); the record follows the sources and sets context.country to AU. No source states a date of occurrence or date of detection, so event.date_occurred and event.date_detected are omitted. The year 2025 for the 22 August disclosure is inferred from the article date. The 'six days later' date is not computed. Source 0 says Deloitte was required to partially refund $440,000, while source 1 says the final instalment of that contract was withheld and DEWR requested $97,587.11; these are not reconciled. Amounts are in dollars as stated and not converted to USD, so impact.direct_loss_usd is omitted. No autonomous agent is described, so event.agent_involved, failure_class and related fields are omitted. event.trigger is omitted because no closed-list cause is stated. Judgements made: responsible_party set to deployer, root_cause set to oversight, severity set to moderate on the bounded-scope test, the risk list limited to RK14, and I7 chosen as the control added by the manual re-review, which the sources do not name. Controls were not reported as held because no source states that. The sources do not say whether a fact-check step existed. [case created by the backfill: read by claude-haiku-5-5, backfill-2026-11-01.2, 2026-11-01; not yet reviewed] [2026-11-01: evidence level set: Preliminary (1 report: HRD (Human Resources Director) · subject to change)]"},"context":{"industry":null,"org_size_band":null,"country":"AU","denominator_available":false,"denominator":null,"related_incident_ids":[],"systemic_exposure":null,"shared_model_id":null,"upstream_incident_id":null,"incident_evidence":null,"regulatory":null,"coverage":null},"narrative":{"steps":[{"step":"went_wrong","text":"The department wrote that Deloitte confirmed these tools likely contributed to the errors in the report.","quote":"You confirm these tools likely contributed to the errors identified in the report.","source_index":1,"grade":"C"},{"step":"reached","text":"The report reached the government with mistakes, including made-up citations.","quote":"Last month, Deloitte was required to partially refund the Australian government $440,000 after it admitted to using AI to help produce a report for the government, which ended up being full of mistakes, including made-up citations.","source_index":0,"grade":"C"},{"step":"caught","text":"Welfare academic Chris Rudge of the University of Sydney first identified the errors, including over a dozen fabricated points.","quote":"The errors were first picked up by welfare academic Chris Rudge from the University of Sydney, who identified over a dozen fabricated points attributed to his colleagues throughout the report.","source_index":0,"grade":"C"},{"step":"done","text":"DEWR requested repayment of $97,587.11, citing the issues with the final product and the failure to follow Deloitte's internal policy on telling the client about AI use.","quote":"In light of the errors and the disclosure breach, the DEWR deputy secretary requested repayment of $97,587.11 “given the issues identified with the final product, and your acknowledgement that internal Deloitte policy regarding communicating the use of AI to the client was not followed”.","source_index":1,"grade":"C"}],"model":"claude-haiku-5-5","prompt_version":"backfill-2026-11-01.2","extracted_at":"2026-11-01","source_urls":["https://www.accountingtimes.com.au/technology/deloitte-ai-debacle-underscores-need-for-government-contract-transparency-ocp-says","https://www.hcamag.com/au/specialisation/hr-technology/deloittes-ai-mess-just-got-worse-and-hr-should-be-paying-close-attention/556535"]},"analysis":{"sections":{"timeline":[{"text":"Deloitte's assurance was later undercut by discovery of a fabricated quote from a leading case on robo-debt.","quote":"That was later undercut by the discovery of a fabricated quote from a leading case on robo-debt.","source_index":1,"when":null},{"text":"A DEWR deputy secretary contradicted Deloitte's account, stating the department was never advised that generative AI caused the errors.","quote":"Six days later, a DEWR deputy secretary contradicted that version of events, writing back that “while Deloitte staff immediately apologised to the department for the errors, and confirmed they were limited to poorly transcribed citations and references, at no point was the department advised that the cause of the errors was attributed to the use of generative AI tools”.","source_index":1,"when":null},{"text":"DEWR requested repayment of $97,587.11.","quote":"In light of the errors and the disclosure breach, the DEWR deputy secretary requested repayment of $97,587.11 “given the issues identified with the final product, and your acknowledgement that internal Deloitte policy regarding communicating the use of AI to the client was not followed”.","source_index":1,"when":null}]},"model":"claude-haiku-5-5","prompt_version":"backfill-2026-11-01.2","extracted_at":"2026-11-01","source_urls":["https://www.accountingtimes.com.au/technology/deloitte-ai-debacle-underscores-need-for-government-contract-transparency-ocp-says","https://www.hcamag.com/au/specialisation/hr-technology/deloittes-ai-mess-just-got-worse-and-hr-should-be-paying-close-attention/556535"]}}