Intellectual Property and Copyright
Training Data Disputes: Major cases like NYT v. OpenAI and Getty v. Stability AI have entered decisive phases to determine if training on copyrighted data is “fair use”.
Output-Based Claims: Newer litigation, such as Disney v. Midjourney, is shifting focus from the training process to whether the AI’s actual outputs (like generated videos or images) infringe on specific creative works.
Largest Settlements: Recent landmarks include a record-breaking settlement by Anthropic with book authors over the use of pirated datasets. 

Autonomous Agent and Liability
Contractual Failures: As “agentic AI” moves from chatbots to autonomous assistants that can sign contracts or execute code, courts are testing who is liable—the user or the developer—when an AI agent makes a costly error.
Product Liability: New standards, including the EU’s Revised Product Liability Directive, now explicitly classify software as a “product,” shifting the burden of proof to developers to show their systems were not defective if they caused harm. 

Consumer Protection and Public Harm
Social Media and Gaming Addiction: Building on precedents from the e-cigarette industry, new lawsuits are testing if “algorithmic addiction” qualifies as a valid legal injury.
Harm to Minors: High-profile cases involving “companion chatbots” have alleged that these systems caused psychological harm or even suicide, leading to significant settlements and new state warnings regarding child safety.
“AI Washing” and Fraud: The SEC and FTC are heavily litigating against companies that overstate their AI capabilities—a practice known as “AI washing”—to mislead investors or consumers. 

Algorithmic Product Liability (The “Public Harm” Frontier) 

Attorneys are successfully arguing that AI software should be treated as a “product” rather than a “service,” making developers strictly liable for physical or psychological harm. 

• Mental Health & Suicide: High-stakes cases, including those against companion chatbot developers, argue that inadequate safeguards led to self-harm or suicide.
• Addictive Design: Similar to the opioid or e-cigarette litigation, class actions are challenging social media and gaming companies for using “algorithmic addiction” to exploit minors, leading to severe public health crises.
• Physical Injury: Cases like Brookes v. Lyft are testing whether digital applications that distract or mislead users during high-risk tasks (like driving) are inherently defective product

Privacy & Automated Deception

New state mandates, particularly in California and Texas, have empowered regulators to treat data misuse as a core consumer harm. 

• Non-Consensual Training: Large settlements are being reached in cases where private data—like voice recordings from smart speakers or home security footage—was used to train AI models without explicit user consent.
• Impersonation Fraud: The explosion of AI-generated deepfakes has led to over 38 states enacting laws that allow consumers to sue for reputational or financial harm caused by digital impersonati