Google's Bard, Gemini, and Gemma AI Systems Allegedly Generated Defamatory Claims About Activist Robby Starbuck, Prompting Lawsuit

Conservative activist Robby Starbuck reportedly filed a defamation lawsuit against Google, alleging its Bard, Gemini, and Gemma AI systems generated false statements linking him to sexual assault and extremist activity. Starbuck claims the outputs caused reputational harm and sought over $15 million in damages. Google acknowledged the hallucinations as known LLM behavior and stated it works to minimize such inaccuracies. See also Incident 1247.

Risk classification

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

The incident involves an AI system generating false or misleading information (defamatory claims about Starbuck), which directly caused reputational and potential financial harm. This aligns with the definition of 'false or misleading information' in the taxonomy.

Additional risk subdomains

• 1.1 Unfair discrimination and misrepresentation: The AI's false claims about Starbuck's ties to white nationalism and sexual assault allegations represent unfair misrepresentation of an individual, causing reputational harm.

Causal factors

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

The incident was caused by Google's AI tools (Bard and Gemini) generating false information unintentionally, as acknowledged by Google's spokesman. The harm occurred after the AI systems were deployed.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

The incident involves an AI system (chatbot) generating false or misleading information, which falls under 'Limited Risk' as defined by the EU AI Act. The report explicitly describes the AI tool producing defamatory content, aligning with the example of 'AI-generated content such as deepfakes' in the Limited Risk category.

AI system and alleged parties

• AI system: Bard, Gemini
• AI purpose: Question Answering; Writing Assistant
• Behaviour type: Assistant
• Alleged developer: Google
• Alleged deployer: Google
• Alleged harmed parties: Robby Starbuck, Epistemic integrity

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Substantial. 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 Substantial, 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 Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: Yes

Directly caused: Starbuck is seeking over $15 million in damages from Google, indicating significant financial and reputational harm caused by the AI's false claims.

Inferred additional harm: The reputational damage could lead to further financial losses, such as lost professional opportunities or reduced influence, though this is not quantified in the report.

Epistemic

Reported: Yes

Directly caused: The AI-generated false claims about Starbuck represent epistemic harm, as they spread misinformation and undermine trust in the accuracy of AI-generated content.

Inferred additional harm: The incident could contribute to broader erosion of trust in AI systems and the information they produce, particularly if similar cases become common.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 1

Potential causes

Management

• Reliance on Disclaimers: Management relies on user warnings rather than ensuring absolute accuracy.

Technology

• LLM Hallucination Tendency: LLMs naturally generate inaccurate but plausible-sounding falsehoods.
• Generation of False Sources: The open model listed fabricated media links to back up false claims.

Data Inputs

• Unverified Training Data: Models trained on unverified web data associate names with negative events.

Human Factors

• Adversarial User Prompting: Users can creatively prompt chatbots to generate misleading statements.

Process and Methods

• Inadequate Inaccuracy Mitigation: Processes failed to prevent newer AI tools from repeating old falsehoods.
• Ineffective Legal Engagement: Google failed to actively engage after receiving cease-and-desist letters.

Regulatory Environment

• Lack of Defamation Precedent: Courts have not established clear liability guidelines for AI defamation.

Information quality

• Classification confidence: Medium
• Reason for confidence: The report provides clear details about the incident, including the AI tools involved, the false claims generated, and the legal action taken. However, some ambiguities remain, such as the exact technical mechanisms behind the AI's errors and the full extent of the harm caused. The report also lacks specific details about the AI's training data or validation processes, which could affect the classification of the incident's causes.
• Ambiguities identified: The exact technical cause of the AI's false claims is not explained; the report does not specify whether the AI's errors were due to training data, algorithmic limitations, or user prompts. The full extent of the reputational and financial harm is not quantified beyond the $15 million lawsuit.
• Alternative interpretations: The incident could potentially be classified under '1.2 Exposure to toxic content' if the false claims are considered toxic, but the primary harm is epistemic (misinformation) rather than the offensiveness of the content itself.
• Missing information: Technical details about the AI's training data, error validation processes, and the specific prompts used to generate the false claims would improve classification confidence. Additionally, more information about the scale of exposure to the false claims would help assess the full extent of the harm.

A civil defamation lawsuit filed by a conservative activist against Google over false AI-generated claims of white nationalist ties and sexual assault. The incident represents a standard commercial LLM hallucination issue with negligible implications for national security.

• Overall national security impact: Negligible
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The incident is a legal dispute regarding past AI outputs and does not present an active, time-sensitive national security crisis.
• Autonomy: Human-controlled. The AI tool functioned as a writing and question-answering assistant, generating text directly in response to human prompts.
• Novelty: Established threat. LLM hallucinations and the generation of false information are well-documented and established limitations of current generative AI technology.

Impact by dimension

• Physical security: Negligible. No physical systems, infrastructure, or human safety were threatened or impacted by the AI-generated false statements.
• Information security: Negligible. While the incident involves AI-generated falsehoods, it represents accidental LLM hallucinations rather than a coordinated state-sponsored information warfare or intelligence compromise.
• Sovereignty: Negligible. No threat to state authority, territorial control, or core government operations; the incident is a civil defamation lawsuit involving a private citizen.
• Economic security: Negligible. The financial impact is limited to a private civil lawsuit seeking damages, posing no threat to national economic stability or strategic technological advantages.
• Societal stability: Negligible. The harm is localized to reputational damage of a single individual, with no evidence of large-scale societal instability, systemic discrimination, or mass civil rights violations.