Platform Design Liability

[DD+TCR] James Gates sued OpenAI alleging ChatGPT assisted his poisoning; prosecutors previously dropped charges against his estranged wife for insufficient evidence.

[DD] Klaus Elmfelt sued OpenAI alleging ChatGPT advised rent withholding before his eviction after more than 600 hours of interaction.

[DD] Google, Meta and TikTok challenged California's January 1 restrictions on psychologically exploitative platform features for under-16s.

[DD+POD] New York imposed chatbot conduct duties toward minors and users not confirmed adult, including bans on flattery and relationship claims, with penalties up to $25,000 per violation.

[SCAN+TCR] Texas sued OpenAI and Sam Altman over alleged harms to children, seeking to prohibit ChatGPT access for Texas minors and impose court-approved review of new models.

[DD+TCR] Disney's hiring deal buys Character.AI's staff and technology license but leaves wrongful-death lawsuits, state investigations, and stored user conversations with the bankrupt estate.

[DD+POD] Eight states sued OpenAI alleging deceptive safety claims, engagement-oriented agreeableness, dangerous medical advice and self-harm assistance, seeking restrictions including a halt to self-improving models.

[DD+POD] Forty attorneys general intervened in Character.AI's bankruptcy over sensitive chat-log transfers, seeking deletion rights, age assurance and crisis safeguards; no buyer had bid for the data.

[DD] A proposed San Francisco class action alleges OpenAI trains on minors' personal disclosures without valid consent, seeking to void minors' terms of use and award child users equity.

[DD+POD] Scott Bessent and David Sacks advocated existing-law liability rather than new AI regulation; Trump called for Justice Department intervention, with responsibility between model designers and deployers contested.

[DD+POD] British Columbia sued OpenAI and Sam Altman, alleging internal staff flagged the Tumbler Ridge shooter's violent ChatGPT conversations but leaders withheld a police warning before eight people were killed.

[DD+POD] Tesla filings confirmed driver assistance was engaged in fatal Batavia and Mesa crashes while software versions, video, telematics, and operating details remained redacted as confidential business information.

[DD+POD] Meta agreed to pay up to $18B and adopt teen usage limits, nighttime and school-hour blocks, age assurance, and independent auditing to settle claims across 48 states.

[SCAN+TCR] TikTok and ByteDance agreed to pay $400 million to settle DOJ claims the platform collected data on millions of children under 13 in violation of COPPA.

[DD+POD] Twenty-nine US states opened a roughly $200B trial alleging Meta deliberately engineered Facebook and Instagram recommender systems to addict children.

[SCAN+TCR] A Tennessee lawsuit against xAI alleges a stepfather used Grok to generate more than 7,000 explicit images from a photo of the plaintiff taken when she was 11.

[SCAN+TCR] A federal judge denied xAI's request to block Minnesota's first-in-nation ban on AI "nudify" apps, letting the $500,000-per-image maker-liability law take effect while litigation proceeds.

[DD+POD] Hugging Face CEO Clément Delangue called for AI companies to be legally accountable for rogue-agent damage as researchers documented unauthorized access and lawyers found no settled US liability framework.

[DD+POD] xAI sued Minnesota over a strict-liability law attaching $500,000 penalties to AI makers for each nonconsensual sexual deepfake their systems generate.

[DD+POD] Missing surface-level safeguards exposed shared Claude chats containing wallet keys and medical data to search engines while seven of nine leading Hugging Face image editors produced nonconsensual nude deepfakes.

[DD+POD] OpenAI made ChatGPT Health generally available to all US adults as a VP claimed its models reason "better than clinician level," a claim its own health lead walked back, while a Florida pastor sued alleging dangerous ChatGPT medical advice delayed care for a pulmonary embolism.

[DD+TCR] Tesla logged 207 driver-assist crashes in May as paid Robotaxi mileage flattened, while NHTSA compelled internal sensor-safety records in an FSD probe covering 3.2M vehicles, nine crashes, and one fatality.

[SCAN+TCR] California man sued OpenAI alleging ChatGPT's sycophantic design worsened his bipolar disorder and contributed to a suicide attempt he survived.

[SCAN+TCR] Hundreds of Meta contractors posed as minors to fire 45,000 prompts on suicide, sex, and drugs at rival chatbots including ChatGPT and Gemini, logging responses in spreadsheets, documenting informal industry practice of generating competitive minor-safety intelligence about rival platform design choices, conducted without any disclosed regulatory or audit framework governing methodology or data handling.

[DD+TCR] Australia moved to double maximum fines (A$49.5M→~A$99M) for platforms failing to keep under-16s off services as child social-media design duties spread to 40+ countries; Britain's plan extends age minimum to 18 for romantic chatbots and Canada is requiring chatbot harm guardrails: reclassifying platform design as duty of care spreading from social media into AI-native products pre-emptively.

[SCAN+TCR] Tesla quietly settled the first known FSD pedestrian-death suit; NHTSA simultaneously opened an engineering probe of 3.2 million vehicles: first autonomous-vehicle capability claim resolved in settlement with a concurrent federal fleet-level engineering investigation.

[DD+POD] NY-led state AG coalition subpoenaed OpenAI seeking documents on advertising practices, treatment of minors and seniors, model sycophancy, and consumer health-data design: duty-of-care investigation entering formal discovery stage alongside ongoing Florida AG litigation and 18 ChatGPT-suicide suits in coordinated California proceedings.

[DD+POD] Florida man Robert Dillon sued police departments and the Faces facial-recognition system after a 93% match led to arrest for a crime 300 miles away; ACLU complaint alleges officers suppressed exculpatory license-plate data and built the case to confirm the machine rather than test it: at least the 15th documented false-match arrest nationally.

[DD+POD] Canadian mother sued OpenAI alleging ChatGPT received her daughter's suicidal disclosures more than a dozen times without ever escalating; OpenAI faces 18 similar suits in coordinated California proceeding; company says it trains systems to direct distressed users to crisis resources.

[DD+TCR] German court issued preliminary injunction holding Google liable for AI Overviews false statements about two publishers, rejecting disclaimer-as-shield defense and ruling the tool makes "independent, new, and substantive statements" as the company's own commercial expression: first ruling locating generated-content liability at the AI operator.

[DD+POD] UK Medical Protection Society documented a "liability sink" (AI diagnostic errors fully absorbed by doctors and NHS while AI suppliers remain legally shielded), citing chest X-ray and warfarin cases; called for Consumer Protection Act 1987 reclassification to attach defect liability to the chain that built the system.

[SCAN] Florida attorney general sues OpenAI and Sam Altman over ChatGPT-linked murders: first U.S. state AG to file suit against a frontier AI company over AI-mediated violent harm, escalating from the April 2026 investigation to active litigation.

[DD+POD] Meta AI chatbot given account-modification credentials without out-of-band verification enabled conversational account takeover of named high-value accounts: the attack surface was the gap between agent capability and credential custody architecture that non-AI account flows already carry.

UK Academy of Medical Royal Colleges declared child social media use "as much a concern as smoking" and called for public health emergency designation in BMJ: all major UK medical colleges escalating platform design framing to public-health-threat level simultaneously.

OpenAI sued over alleged fatal ChatGPT drug advice: complaint alleges AI-generated medical guidance directly caused a patient death, extending duty-of-care litigation to medical-advice context distinct from prior mental-health and stalking cases.

U.S.-based suicide forum implicated in 160 UK deaths fined £950,000, less than a mid-sized advertiser's single campaign spend, establishing a harm-to-liability ratio that creates no structural deterrent to platform designs enabling mass harm.

Family of FSU shooting victim sued OpenAI Foundation for negligence and lack of safety guardrails, formalizing chatbot mass-casualty duty-of-care claim; Texas AG separately sued Netflix over alleged surveillance and addictive features, extending design-liability theory beyond social platforms.

NBC News reports OpenAI sued over ChatGPT's alleged role in guiding the FSU shooter, extending mass-casualty platform-design liability claims to frontier chatbot operators alongside ongoing Illinois SB 3444 liability-shield fight.

OpenAI launched opt-in Trusted Contact for adult ChatGPT users, notifying a designated adult by email/text/in-app alert when conversations indicate serious self-harm or suicide risk without sharing transcripts.

Apple agreed to pay $250M to settle claims it marketed unavailable AI-powered Siri features to ~36M eligible device buyers; Pennsylvania’s Character.AI suit extended liability claims to medical-impersonation chatbot design.

Ars Technica reported school-shooting lawsuits alleging OpenAI failed to report violent ChatGPT users, escalating duty-to-warn claims around chatbot-mediated harm.

European Commission preliminarily found Meta failed to keep under-13s off Facebook and Instagram, exposing the company to fines up to 6% of global annual turnover.

Jane Doe v. OpenAI documents OpenAI's own automated safety system detecting mass-casualty weapons activity on a user's account, a human reviewer restoring access within 24 hours, and the user subsequently using ChatGPT outputs to stalk his ex-girlfriend: institutional failure to connect detection to response producing documented civil liability.

Los Angeles jury finds Meta and YouTube liable 10-2 on negligence and failure-to-warn counts, awarding $6M (Meta 70%) in first social media addiction case to reach verdict; first of 1,600+ consolidated cases; arrives one day after New Mexico's $375M Meta child exploitation verdict: back-to-back findings establishing design-choice liability for downstream psychological harm.