Google Bard Allegedly Generates False Allegations Against Consulting Firms Used in Research Presented in Australian Parliamentary Inquiry

Australian academics reportedly used Google Bard AI to generate case studies for a parliamentary inquiry, leading to false allegations against major consultancy firms. The AI-generated misinformation prompted an apology from the academics, causing reputational harm for all parties involved and raising concerns about the reliability of AI tools in producing accurate and unbiased information.

Risk classification

• Primary risk domain: 3 Misinformation
• Primary risk subdomain: 3.1 False or misleading information

The AI system generated false and misleading allegations of financial misconduct, which were then published in a public parliamentary submission.

Additional risk subdomains

• 5.1 Overreliance and unsafe use: The academics overrelied on the AI tool, assuming its authoritative-sounding output was accurate without performing basic fact-checking.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The reputational risk was caused by Google Bard generating false allegations, which was an unexpected and unintentional outcome of the academics using the tool to draft their submission.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: Google Bard is a general-purpose chatbot and content generation tool, which falls under the limited risk category requiring transparency about its AI nature.

AI system and alleged parties

• AI system: Google Bard (Google)
• AI purpose: Writing Assistant; Technical Text Generation
• Behaviour type: Assistant
• Alleged developer: Google Bard
• Alleged deployer: James Guthrie
• Alleged harmed parties: Parliament of Australia, KPMG, James Guthrie's co-authors, James Guthrie, Deloitte

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Minor
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Epistemic

Reported: The report explicitly describes epistemic harm in the form of AI-generated false allegations and fabricated case studies.

Directly caused: Google Bard fabricated multiple detailed case studies of financial misconduct, falsely accusing KPMG and Deloitte of specific scandals.

Indirectly caused: These fabricated allegations were published in an official parliamentary submission, temporarily corrupting the public record before being retracted.

People affected

• Occurrences reported: 1
• People reportedly exposed: 4

Potential causes

Management

• Lack of Co-Author Oversight: Co-authors did not review or verify the lead author's AI-generated content.

Technology

• AI Hallucination: Google Bard generated realistic but false case studies of misconduct.
• Authoritative-Sounding Output: The LLM generated highly confident text that masked its factual inaccuracy.

Data Inputs

• Lack of Source Verification: The model lacked real-time verification of historical audit records.

Human Factors

• Overreliance on AI Output: Academics trusted the AI-generated facts without secondary verification.
• Lack of AI Literacy: The lead academic had only started using Google Bard that same week.

Process and Methods

• Inadequate Fact-Checking: The authors failed to verify the generated claims before submission.
• Flawed Submission Drafting: The research process integrated unverified AI outputs directly into draft.

Regulatory Environment

• Abuse of Parliamentary Privilege: Privilege protected false accusations from immediate legal defamation checks.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides clear, direct, and uncontradicted accounts of the incident, including statements from the affected firms, the Senate committee chair, and the academic who took responsibility.

Australian academics unintentionally submitted fabricated case studies generated by Google Bard to a parliamentary inquiry, falsely accusing KPMG and Deloitte of misconduct. The incident highlights minor risks to government functions through the potential pollution of official public records, though it was quickly corrected with no lasting national security impact.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: Australia
• Alleged perpetrator: Australian academics

Threat characteristics

• Imminence: Long-term. This represents an ongoing strategic concern regarding the verification of AI-generated content in official processes, rather than an active crisis.
• Autonomy: Human-controlled. The AI tool was used as a writing assistant, and humans retained final control over submitting the generated text to the inquiry.
• Novelty: Established threat. LLM hallucinations and the generation of authoritative-sounding false information are well-documented and established behaviors of these models.

Impact by dimension

• Physical security: Negligible. No kinetic threat, physical damage, or risk to human safety was reported in connection with this incident.
• Information security: Negligible. No evidence of coordinated foreign influence or state-sponsored disinformation; the incident was an unintentional result of AI hallucination.
• Sovereignty: Minor. The incident temporarily introduced fabricated evidence into an official Australian parliamentary inquiry, briefly compromising the integrity of the public record before being corrected.
• Economic security: Negligible. While the false allegations targeted major financial consultancy firms, there was no systemic threat to financial stability or strategic technology theft.
• Societal stability: Negligible. No threats to social cohesion, civil liberties, or population safety occurred as a result of this localized academic error.