OpenAI Alleged by Lawsuit Violated Users' Privacy Rights by Training AI on Private Info without Informed Consent

OpenAI's products such as ChatGPT and DALL-E were alleged in a lawsuit using stolen private information from internet users without their informed consent or knowledge.

Risk classification

• Primary risk domain: 2 Privacy & Security
• Primary risk subdomain: 2.1 Compromise of privacy by obtaining, leaking or correctly inferring sensitive information

The lawsuit alleges that OpenAI unlawfully scraped and obtained private, non-public sensitive information, including medical data and private conversations, to train its models without consent.

Additional risk subdomains

• 6.3 Economic and cultural devaluation of human effort: The lawsuit claims OpenAI used copyrighted works without compensation, threatening professional obsolescence and devaluing human effort.
• 6.5 Governance failure: The reports highlight the lack of federal AI regulations in the US, forcing reliance on class-action lawsuits to establish safety guardrails.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Pre-deployment

The data scraping and training practices were human decisions made by OpenAI's management and developers prior to the deployment of the models.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The systems involved are primarily chatbots (ChatGPT) and image generators (DALL-E), which are explicitly classified as limited risk under the EU AI Act, requiring transparency obligations.

AI system and alleged parties

• AI system: ChatGPT, Dall-E (OpenAI)
• AI purpose: Chatbot; Visual Art Generation
• Behaviour type: Assistant
• Alleged developer: OpenAI
• Alleged deployer: OpenAI
• Alleged harmed parties: social media users, internet users, Children

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Severe. 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 Severe, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Severe, indirect Severe
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Civil rights

Reported: Yes, the report explicitly describes violations of privacy and property rights as a central claim of the class-action lawsuit.

Directly caused: The lawsuit alleges that OpenAI violated the privacy and civil rights of hundreds of millions of internet users, including children, by scraping their personal data without consent.

Inferred additional harm: The widespread scraping of personal data without consent likely undermines systemic privacy rights and digital autonomy for internet users globally.

Privacy

Reported: Yes, the report explicitly describes massive privacy violations caused directly by the scraping of personal data.

Directly caused: OpenAI allegedly scraped personally identifiable information, private conversations, medical data, and children's information from hundreds of millions of internet users without consent.

Indirectly caused: Integrating these models into third-party apps like Snapchat, Slack, and MyChart allows real-time capture of user locations, financial transactions, and private health information.

Inferred additional harm: It is highly likely that millions of additional users have had their private data ingested and processed without their knowledge, leading to ongoing, unquantified privacy violations.

People affected

• Occurrences reported: 1
• People reportedly harmed: 200000000
• People reportedly exposed: 200000000

Potential causes

Management

• Profit Prioritization over Ethics: Restructuring to a for-profit model, abandoning original safety principles.
• Reckless AI Arms Race: Rushing products to market to beat competitors without testing safety risks.
• Creating Economic Dependency: Embedding untested AI into society before understanding correlation failures.

Technology

• Lack of Safeguards and Controls: Products rushed to market without controls to stop harmful outputs.
• Opaque Black Box Models: Immature technology whose internal workings and alignment are not understood.
• Rapid API and Plug-in Integration: Embedding immature models into critical third-party apps and infrastructure.

Data Inputs

• Mass Scraping Without Consent: Scraping 300 billion words of private social media data without permission.
• Misappropriation of Child Data: Collecting personal info of minors without informed parent consent.
• Lack of Data Opt-Out Mechanisms: Initially no way for users to opt out of their data being used for training.

Human Factors

• User Trust in Unsecured Systems: Users inputting sensitive personal, financial, and medical data into chatbots.
• Underage User Engagement: Children under 13 using the tools without proper verification or safeguards.
• User Exposure to Defamation: Chatbots producing false information that wrongfully accuses users of fraud.

Process and Methods

• Secret Harvesting of Web Data: Harvesting massive amounts of personal data secretly without notice to owners.
• Deficient User Verification: Failing to verify user age or ensure children under 13 are not using tools.
• Opaque Privacy Policies: Lack of clear, conspicuous disclosure on what data is collected and stored.

Regulatory Environment

• Lack of US Data Regulations: Absence of clear legal distinctions between public domain and free-to-use data.
• Failure to Register as Data Broker: OpenAI did not register as a data broker as required under applicable laws.
• Lack of Federal AI Guardrails: Government is a long way from being AI-ready with adequate federal oversight.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, detailed information about the class-action lawsuit, the specific legal claims (privacy violations, negligence, larceny), the AI models involved (ChatGPT, DALL-E, Vall-E), and the nature of the data scraping. There is high consensus across the sources regarding the facts of the filing.
• Ambiguities identified: The exact volume of non-public vs. public data scraped is legally debated, as is the definition of 'stolen' public data under US law.
• Alternative interpretations: The defense argues that scraping public internet data constitutes 'fair use' under copyright law, presenting an alternative legal interpretation of the same facts.

A class-action lawsuit filed against OpenAI highlights significant privacy and data governance concerns regarding the unauthorized scraping of personal data to train AI models. While the scale of affected individuals is vast, the national security implications are minor and currently managed through domestic legal and regulatory channels.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: OpenAI

Threat characteristics

• Imminence: Long-term. The incident involves a class-action lawsuit addressing past and ongoing data collection practices rather than an active, fast-moving crisis.
• Autonomy: Human-controlled. The data scraping and training decisions were entirely planned and executed by human developers and management at OpenAI.
• Novelty: Evolved capability. Web scraping is an established practice, but the scale and application to train massive commercial foundation models represents an evolved capability.

Impact by dimension

• Physical security: Negligible. No physical systems, kinetic weapons, or critical infrastructure were targeted or affected by the data scraping practices.
• Information security: Minor. While the scraping included private conversations and medical data, there is no evidence of state-sponsored intelligence compromise or active information warfare operations.
• Sovereignty: Negligible. The incident is a domestic civil lawsuit and does not threaten state authority, border control, or core government operations.
• Economic security: Minor. The lawsuit alleges unauthorized commercial use of data and potential professional obsolescence, representing minor economic friction but no strategic technology theft.
• Societal stability: Minor. Widespread collection of personal data without consent raises significant privacy concerns for millions, but is being addressed through standard legal channels.