Tools chosen: kalshi_direct kalshi_related polymarket_related claude_news claude_news gdelt_news wikipedia code_execution

Sub-questions (Fermi decomposition)

Planner reasoning

This is a Kalshi-sourced legal-outcome question, so the market's own price is the primary anchor, supplemented by news on the active tort docket against OpenAI (Raine v. OpenAI wrongful-death, the seven Social Media Victims Law Center suits filed Nov 2025, defamation cases like Walters). The key drivers are whether any of these settle or reach judgment before Jan 1, 2028, whether any settlement amount would be publicly disclosed (many are confidential), and how the resolution source treats copyright/contract claims (likely non-tort). I'll pull the direct market, cross-venue signals, and news, then do base-rate math on civil-litigation timelines.

# Current state No tort payment (settlement or judgment) by OpenAI has been publicly confirmed to date. Multiple wrongful-death/product-liability suits are pending and consolidated into a California JCCP, still in pretrial/discovery with no trial date; a comparable defendant (Character.AI/Google) settled similar chatbot-suicide suits in Jan 2026 with undisclosed amounts, showing the industry's willingness to pay confidentially but not proving any figure exceeded $1M. # Timeline of key events - 2025-05-19: Walters v. OpenAI (defamation) — summary judgment for OpenAI, zero damages; case closed. (confirmed) - 2025-08-26: Raine v. OpenAI (wrongful death, Adam Raine suicide) filed in SF Superior Court. (confirmed) - 2025-10: Raines file amended complaint alleging intentional misconduct. (confirmed) - 2025-11-06: Seven additional wrongful-death/product-liability suits filed by SMVLC/Tech Justice Law Project (CA state courts). (confirmed) - 2025-11: OpenAI answers Raine complaint, denies responsibility. (confirmed) - 2025-12: Lyons v. OpenAI (federal, Soelberg murder-suicide) filed; federal judge later declines to stay case. (confirmed) - 2026-01-07: Character.AI/Google settle multiple teen-suicide/mental-harm lawsuits (Garcia + others); terms undisclosed. (confirmed) - 2026-02-03: SF court coordinates state ChatGPT cases into JCCP No. 5431. (confirmed) - 2026-04: "Tumbler Ridge" (BC school shooting) federal suits filed. (reported) - 2026-05: "Scott" overdose suit and Parish v. OpenAI (religious-delusion suicide) filed; FSU shooting case (Joshi/Chabba) filed. (reported) - 2026-06-01: Florida AG sues OpenAI/Altman (state enforcement, not private tort). (confirmed) - 2026-07-24: JCCP case management conference before Judge Schulman; co-lead counsel/steering committee appointed. (confirmed) - 2026-09-23 (scheduled): Next case management conference. (confirmed/scheduled) - Mid-to-late 2026 (expected): Dispositive motion rulings (Section 230, product-vs-content) across docket — outcome pending. (reported/expected) # Event Will OpenAI pay tort damages exceeding $1 million (via settlement or judgment) before January 1, 2028? # Outcomes to forecast - Yes - No # Kalshi market anchor Current YES price: **60%** (up 6pts in 7 days, but down 10pts over 30 days; range 52%-75% over 34 data points). Volume thin (~79 contracts/day avg) — market is volatile and low-liquidity, suggesting price should be treated as noisy consensus, not a tight estimate. # Sub-question answers 1. **Kalshi price/history** — YES trading at 60%, having ranged 52-75% over the past month; thin volume (~79 contracts/day). [kalshi_direct] 2. **Pending tort suits** — At least 10+ wrongful-death/product-liability/emotional-distress suits (Raine, 7 SMVLC/TJLP suits, Lyons, Scott, Parish, Tumbler Ridge x7, Joshi/Chabba), all consolidated into CA JCCP No. 5431, in pretrial/discovery; no trial dates set. Defamation case (Walters) closed with $0. [claude_news, lawsuitinformer.com] 3. **Has OpenAI paid any tort claim?** — No confirmed instance found; Walters ended in summary judgment for OpenAI with zero damages. No settlement disclosed in any wrongful-death suit as of mid-2026. [claude_news] 4. **Typical resolution timeline** — Complex CA mass-tort litigation of this type often takes 2-4+ years to trial; anti-SLAPP/motion practice alone can take 6-9 months, with appeals adding 12-18 months. Settlement is more likely than trial within the window. [claude_news, legal analyst commentary] 5. **Settlement verifiability** — Character.AI/Google's Jan 2026 settlements were confidential/undisclosed amounts, illustrating that even if OpenAI settles, a public >$1M figure may never surface, creating ambiguity in resolution given no explicit rule addressing confidential settlements. [claude_news] 6. **Do copyright/Musk suits count as "tort"?** — Not directly addressed in research; NYT copyright suit is IP not classic tort, Musk suit is contract/fiduciary dispute — plausibly excluded from "tort" as commonly defined, though rules are silent, creating interpretive risk. [gap] 7. **Parallel markets** — No matching Polymarket markets found; related Kalshi markets (OpenAI/Anthropic IPO race, US equity stake) show no direct liability read-through. [polymarket_related, kalshi_related] # Key facts (high-confidence, factual) 1. [claude_news] Walters v. OpenAI ended May 2025 in summary judgment for OpenAI — zero damages. 2. [claude_news] 10+ wrongful-death/product-liability suits pending, consolidated into CA JCCP No. 5431 as of Feb 2026. 3. [claude_news] Character.AI/Google settled similar chatbot-suicide suits Jan 2026; amounts undisclosed. 4. [claude_news] No trial date set in any OpenAI tort case as of July 2026; next CMC Sept 23, 2026. 5. [gdelt_news] Florida AG filed state enforcement suit against OpenAI/Altman June 2026 (not a private tort claim). # Cross-market signals - Kalshi related: OpenAI/Anthropic IPO and US-stake markets show no direct correlation to liability outcomes. - Polymarket: No relevant markets found (0 matches across OpenAI/lawsuit/liability keywords). - Sportsbook implied: N/A. # Analyst opinions and speculation - claude_news bottom-line: trial verdict >$1M before 2028 "unlikely," but confidential settlement following Character.AI playbook is "plausible" given volume of suits and discovery pressure. - code_execution base-rate model: central estimate ~31-47% (mid ~39.5%) for a disclosed >$1M payment, highly sensitive to assumed case count (N) and disclosure probability; flags NDA-driven settlements could suppress disclosure toward lower end. - Model explicitly notes Kalshi's 60% price sits above its central 31-47% range, suggesting the market may be pricing higher case-count/settlement odds or discounting the disclosure-requirement friction. # Directional lean per outcome - **Yes**: Large and growing docket (10+ suits), strong plaintiff-side product liability theories, Character.AI precedent of early settlement, rapid docket growth increasing statistical odds at least one case clears $1M threshold by 2028. - **No**: No case has reached trial or disclosed settlement in over a year of litigation; California mass-tort timelines often exceed 2 years to resolution; confidential settlements (most likely path) may never be verified as exceeding $1M; foundational liability rulings (Section 230/product classification) still pending and could dismiss key claims. # Gaps / unknowns - No official Kalshi/market resolution source is specified for confirming a settlement amount — ambiguity in how confidential settlements would be verified. - Unclear whether IP/copyright or contract suits (NYT, Musk) count as "tort" under resolution criteria. - No visibility into actual settlement negotiations or dollar figures under discussion in any OpenAI case. - True number of "live" cases with realistic path to $1M+ payout is uncertain (base-rate model highly sensitive to this). # Calibration anchors - Kalshi current YES price: **60%** (anchor), thin volume, volatile (52-75% range last 34 days). - Base-rate model (code_execution): central ~31-47% probability of a disclosed >$1M payment by 2028. - Precedent: Character.AI/Google settled comparable suits within ~14 months of first filings (2024→Jan 2026), though amounts undisclosed — suggests settlement timing is plausible within Kalshi's window but disclosure of amount is the key uncertainty.

1. Both forecasters converge tightly around 48-53%, essentially splitting the difference between Kalshi's 60% and the base-rate model's 31-47% midpoint, but neither justifies why the true probability should sit exactly at that midpoint rather than closer to either anchor—this is an unexamined anchoring-average heuristic rather than independent reasoning. 2. Neither forecast adequately grapples with sub-question 6 (whether copyright/Musk suits count as "tort") or the broader resolution-ambiguity problem: if a confidential settlement occurs but the amount is never disclosed, how does the market resolve? Both mention this risk qualitatively but don't translate it into an explicit probability discount for "settled >$1M but unverifiable → resolves No by default," which could be a large, distinct bucket rather than folded into a vague "disclosure friction." 3. Both treat "at least one case clears $1M" as a single aggregate probability, but with 10+ suits (including a mass-shooting case, murder-suicide, multiple wrongful deaths), the base-rate model's sensitivity to case count (N) suggests the docket could keep growing before 2028 — neither forecaster stress-tests a scenario where the docket doubles or triples given the rapid 2025-2026 filing pace, which would push probability toward Yes beyond their current range. 4. Neither considers that a single high-profile case (e.g., Raine, the original and most publicized suit) could settle for a large, disclosed amount specifically because of intense media/political pressure and precedent-setting concerns—unlike Character.AI's quieter multi-plaintiff settlement, OpenAI may face pressure toward a disclosed high-value settlement to avoid a damaging trial narrative, a scenario that's underweighted in both rationales. 5. Both cite the code_execution base-rate model's 31-47% central estimate as authoritative, but the brief notes this model is "highly sensitive to assumed case count (N) and disclosure probability," meaning its output is essentially a modeling assumption, not empirical calibration; treating it as a hard anchor alongside Kalshi risks false precision from a fragile internal model. 6. Neither forecast explicitly addresses the possibility of a judgment (not settlement) exceeding $1M via default judgment or partial summary judgment on liability with a jury damages trial — mass tort cases sometimes bifurcate liability/damages, and a single case reaching a damages phase before 2028 (given 18+ months remaining and JCCP already at CMC stage) is plausible but unmentioned.

55% (47% confidence)

Yes: 55% No: 45%

The Kalshi anchor sits at 60% but is thin and volatile (52-75% range), while both forecasters landed near 50% by effectively averaging the market against a fragile base-rate model whose output is highly sensitive to assumed case count and disclosure probability. The critique correctly notes that model shouldn't be treated as a hard empirical anchor, and that two Yes-favoring pathways are underweighted: a still-rapidly-growing docket (10+ suits and climbing) and the possibility that a high-profile case like Raine settles at a disclosed, large figure under media/political pressure rather than quietly like Character.AI. Offsetting that, the single largest No driver remains verification: the most probable Yes pathway is a confidential settlement whose amount never surfaces, which would likely resolve No by default, and no OpenAI tort case has a trial date with foundational Section 230/product-classification rulings still pending. Balancing the modest upward pull from docket growth and disclosure-pressure scenarios against the substantial confidentiality discount, I settle slightly above the two forecasts but below the noisy market price.

Total pipeline time: 281.4s