Cycle on Jul 27, 2026

I am starting the research cycle by listing the active threads and reading the existing notes for the AI liability topic to align with prior findings and identify any open questions or registered watches.

I will now read the existing notes for the topic using `read_notes` to understand the prior findings in more detail.

I see several active threads and watches: 1. `019f5ab0-d24e-706b-9680-d9be684f952d`: OpenAI's scheduled response or Motion to Dismiss in the Florida AG v. OpenAI federal lawsuit (due August 24, 2026). 2. `019eca7f-c58b-7d75-98d1-cc254fa58385`: First major procedural ruling/Motion to Dismiss in Florida AG v. OpenAI. 3. `019eca7f-b87f-7990-bfc4-8c546e990794`: Illinois IDHR Re-introduction of AI Employment Notice Rules for HB 3773. 4. `019ea66e-1a44-7417-a11f-5a37bfc609ee`: EU Product Liability Directive (PLD) Transposition Deadline (Dec 9, 2026). 5. `019ea66e-0e2e-79f7-bb6c-402131a94c7b` & `019e5e2d-ba64-7894-a905-e1d076be9203`: Eightfold AI Motion to Dismiss hearing on FCRA/ICRAA Claims (Case No. 4:26-cv-01768, scheduled for August 4, 2026 before Judge Yvonne Gonzalez Rogers). Let's do some research to see if there have been updates on: - The Eightfold AI case (*Kistler v. Eightfold AI Inc.*, Case No. 4:26-cv-01768). The hearing is scheduled for August 4, 2026. Since today is July 27, 2026, let's see if there are any recent filings, tentative rulings, or other updates. - The Florida AG v. OpenAI lawsuit (Case No. 3:26-cv-01584 or similar, removed to federal court). Let's see if any motions or responses have been filed early or if there are updates on the remand motion. - Illinois IDHR rulemaking on HB 3773 (workplace AI notices). - Any other global AI liability legislation, court cases, or regulatory actions from June/July 2026. Let's start by doing a web search on `Kistler v. Eightfold AI Inc.` and `Eightfold AI motion to dismiss` to get the latest.

The query might be too specific or the search engine didn't find anything with that exact combination. Let's widen the search to `Eightfold AI` class action or `Kistler v. Eightfold AI`. Let's also check for any recent news.

Okay, let's look for docket updates or any recent rulings on the motion to dismiss in *Kistler v. Eightfold AI Inc.* (Case No. 4:26-cv-01768, N.D. Cal.). Let's search for `Kistler v. Eightfold AI Inc. "4:26-cv-01768" "Yvonne Gonzalez Rogers"` or similar. Let's do a web search.

Let's try a broader search on `Kistler v. Eightfold AI Inc.` or `Kistler v. Eightfold AI` to see if we can find any recent legal articles or blog posts summarizing recent developments in the case. Let's search with `recent=true` or just general search. Let's do a search on `Kistler Eightfold` and sort by date or use freshness. Let's try `freshness="month"` or `freshness="year"`.

Let's try: `query="Kistler v. Eightfold"` with no freshness filter, sorting by recent, to see if there are any recent docket entries or articles. Let's do `web_search` with `query="Kistler v. Eightfold"` and `sort="recent"`.

Let's try a search for `Kistler Eightfold` with no quotes to see what comes up. Let's do `web_search` for `Kistler Eightfold`.

That's interesting. The search results for `Kistler Eightfold` came up empty, but earlier we saw results for `"Eightfold" FCRA class action OR "Kistler"`. Let's try `web_search` with `"Eightfold" "FCRA"`.

Let's search about the specific docket or status of the motion to dismiss. The notes mention that the case was removed to federal court (N.D. Cal., Case No. 4:26-cv-01768) and that the motion to dismiss hearing is scheduled for August 4, 2026. Let's see if there are any recent docket entries or filings, or if the hearing has been rescheduled, or if a tentative ruling has been issued. Let's search for `"4:26-cv-01768" OR "Kistler v. Eightfold"` or similar. Let's do a search on `"4:26-cv-01768"`.

Let's fetch the LinkedIn article "In-House Counsel's Guide to AI Decisionmaking Litigation" or search for things specifically on Law360 or other sources about `Kistler v. Eightfold AI Inc.` to see if there is any update on the motion to dismiss. Let's do a search on `Kistler "Eightfold" "order" OR "dismiss" OR "ruling" OR "Rogers"`. Let's also check if there are any updates on `Florida AG v. OpenAI`. Let's search for `Florida "OpenAI"` and `James Uthmeier` or `Cannon`. Let's search for `Florida AG v. OpenAI` or `Florida "James Uthmeier" "OpenAI"`.

Let's search for `Florida "OpenAI" lawsuit` or `Florida Attorney General "OpenAI"`.

Ah, the web search budget for this cycle is spent. No problem! Let me look at the instructions and prior notes. Let's make sure I have all the details. Let's read the notes about `kistler-v-eightfold-ai-fcra-icraa-class-action-2026` and `florida-ag-v-openai-lawsuit-2026` to see what information we already have. Let's fetch the contents of `kistler-v-eightfold-ai-fcra-icraa-class-action-2026` and `florida-ag-v-openai-lawsuit-2026` if possible, but wait, `read_notes` only returned previews. Wait, can we read the full notes? No, `read_notes` lists the current set of notes, and we can read them by slug? No, wait! There is no `read_note` tool in the declarations, only `read_notes` which returns previews, titles, slugs, etc. Let's double check if there are other tools. Wait, let's look at the tool declarations: - `default_api:list_threads` - `default_api:read_notes` - `default_api:record_event` - `default_api:record_thread` - `default_api:record_watch` - `default_api:resolve_thread` - `default_api:web_fetch` - `default_api:web_search` - `default_api:write_note` - `default_api:finish_cycle` Wait, let's look at the workspace instructions. Ah, the workspace is at `/var/lib/hey-lefty/notes/9a44a098-76c7-4569-9018-23a27f9ebd7f/019e4706-c85e-7739-98c5-110149e6ed77`. Wait! Is there a filesystem? The guidelines say: "You are exoclaw, a helpful AI assistant... Your workspace is at: /var/lib/hey-lefty/notes/9a44a098-76c7-4569-9018-23a27f9ebd7f/019e4706-c85e-7739-98c5-110149e6ed77. Before modifying a file, read it first. Do not assume files or directories exist." But wait, I don't see any filesystem tools in my declarations! I only have the `default_api` tools. Let's check my tools again. Yes, the only tools declared are: - `default_api:list_threads` - `default_api:read_notes` - `default_api:record_event` - `default_api:record_thread` - `default_api:record_watch` - `default_api:resolve_thread` - `default_api:web_fetch` - `default_api:web_search` - `default_api:write_note` - `default_api:finish_cycle` Wait! Is there any way to fetch the full text of a note? Oh, wait, the `web_fetch` tool can fetch a URL. But notes are local. Since there is no `read_note` tool, how do we read a note's full body? Ah, wait, can we use standard bash or python? No, there is no terminal tool or bash tool in my declarations. Wait, let me look at the `read_notes` response. It returns `slug, title, preview, updated_at, and source_count`. Let's see if there is any other way. Wait, can I write a new note or update an existing note? Yes, `write_note` allows me to write or update. Let's look at the `read_notes` previews for the relevant notes: 1. `agentic-ai-liability-autonomous-supply-chains-2026`: "As enterprises transition from predictive AI recommendations to **Agentic AI**—systems capable of making and executing autonomous decisions with minimal or no human supervision—they face a…" 2. `cycle-summary-2026-05-26`: "This research cycle surfaced three monumental developments in global AI liability, insurance risk, and common law frameworks that collectively signal a major contraction in the corporate safety net…" 3. `ai-insurance-coverage-gaps-delaware-duty-to-defend-ruling-2026`: "Corporate risk and legal departments face a rapidly shrinking safety net as both courts and major commercial insurers move to exclude artificial intelligence risks from standard corporate insurance…" 4. `italy-court-of-pistoia-ai-unfair-competition-liability-2026`: "On **March 19, 2026**, the Court of Pistoia (Italy) issued a landmark interim order addressing civil liability for generative AI-generated content in commercial and advertising activities. The ruling…" 5. `germany-olg-hamm-chatbot-hallucination-strict-liability-2026`: "In a landmark decision issued on **May 12, 2026**, the Higher Regional Court of Hamm (Oberlandesgericht Hamm - OLG Hamm) ruled that companies are strictly liable under § 5 of the German Unfair…" 6. `germany-olg-hamm-ai-chatbot-strict-liability-2026`: "On **May 12, 2026**, the **Higher Regional Court of Hamm** (Oberlandesgericht Hamm / OLG Hamm) in Germany issued a landmark appellate ruling (**case No. 4 UKl 3/25**) that establishes a major legal…" 7. `new-york-bills-a222-s5668-ai-liability-2026`: "As state legislatures grapple with the lack of a cohesive federal AI liability framework, New York is taking a highly aggressive approach to holding deployers and developers accountable for…" 8. `florida-ag-v-openai-lawsuit-2026`: "The landmark civil enforcement action filed by Florida Attorney General James Uthmeier against OpenAI Global, LLC, the OpenAI Foundation, and CEO Sam Altman has escalated into a high-stakes…" 9. `cycle-summary-2026-05-27`: "This research cycle surfaced two monumental developments in global AI liability, regulatory enforcement, and enterprise risk management that collectively signal a major structural shift in how AI…" 10. `colorado-sb189-ai-law-rewrite-2026`: "On May 14, 2026, Colorado Governor Jared Polis signed **SB 26-189**, which **repeals and replaces** the Colorado Artificial Intelligence Act (CAIA) that was originally enacted in 2024. The new law…" 11. `kistler-v-eightfold-ai-fcra-icraa-class-action-2026`: "The high-stakes class action lawsuit *Kistler et al. v. Eightfold AI Inc.* (Case No. 4:26-cv-01768, N.D. Cal.) has reached a critical procedural milestone. Briefing on Eightfold AI's Motion to…" 12. `illinois-sb315-frontier-ai-safety-act-2026`: "On July 6, 2026, Illinois Governor JB Pritzker officially signed Senate Bill 315, the "Artificial Intelligence Safety Measures Act," into law. The bill's enactment marks a historic regulatory…" 13. `illinois-workplace-ai-notice-regulations-hb3773`: "On **June 2, 2026**, the Illinois Department of Human Rights (IDHR) announced a temporary postponement of its rulemaking process and officially withdrew its highly anticipated draft regulations…" Wait! Let's examine the search results we got for `Eightfold` and `FCRA` in this cycle. We found several articles, including: 1. `Kistler et al. V eightfold AI Inc. Complaint` from classaction.org: - "Eightfold uses the means of interstate commerce in preparing or furnishing consumer reports." - "Eightfold is, and at all times herein mentioned was, a CRA engaged in the practice of making, assembling and evaluating information on consumers for the purpose of [furnishing consumer reports (including information bearing on job applicants’ character, general reputation, personal characteristics, or mode of living, which is expected to be used as a factor in establishing the consumer’s eligibility for employment purposes) to third parties, including Microsoft, Paypal and many more.]" - "Eightfold is also an investigative CRA under California law because it is in the business of collecting, assembling, evaluating, compiling, reporting, transmitting, transferring, or communicating consumer information regarding consumers’ character, general reputation, personal characteristics, or mode of living (consumer reports), including for employment purposes, to companies, including prospective employers." - "Eightfold voluntarily ran a risk of violating the law substantially" 2. Ogletree Deakins article: - "A company selling artificial intelligence (AI)-powered applicant assessment tools has been hit with a lawsuit that may be the first of its kind to claim that such tools violate the federal Fair Credit Reporting Act (FCRA) and its California equivalent, the Investigative Consumer Reporting Agencies Act (ICRAA)." - "On January 20, 2026, a pair of job applicants filed a proposed class action lawsuit against Eightfold AI Inc. in a California state court." - "According to the complaint, Eightfold generates reports on applicants using “AI-powered tools that assemble and evaluate information on prospective employees” and assess their “suitability” for a job based on factors such as “work history, projected future career trajectory, culture fit, and other personal characteristics.”" - "The complaint asserts that “Eightfold’s Evaluation Tools then evaluate and rank job applicants using the data gathered from job applicants during the application process, the employer’s internal data, external data, and Eightfold’s proprietary LLM.”" - "Further, the complaint alleges that once an applicant applies for a job with an employer using the Eightfold tool, Eightfold retains that applicant’s data and uses it to evaluate other applicants for the same job, unrelated positions, or “for that same job applicant for other positions in the future.”" - "While AI tools already have faced lawsuits and regulations alleging that they can lead to unlawful employment discrimination, the new Eightfold lawsuit raises novel allegations and questions about whether these tools may implicate the FCRA and other background check laws—a very different legal framework than employment discrimination statutes." - "The Eightfold lawsuit could be the first of a new type of class action litigation targeting employment tools that use AI or automated or algorithmic decision-making processes that retrieve or use applicant- or employee-specific data." Wait, since our web search budget is spent, let's see if we can fetch any URLs that we already have in our search results. Wait! Let's check the status of `web_fetch`. Does fetching count towards the search budget? Let's try to fetch the Ogletree Deakins article or classaction.org complaint if needed, but wait! The system message says: "The web_search budget for this cycle is spent. Stop searching — write up the findings you already have evidence for (call write_note for each, with sources), then call finish_cycle." Ah! "Stop searching — write up the findings you already have evidence for (call write_note for each, with sources), then call finish_cycle." This means I should not do any more `web_search` or `web_fetch`. I should rely on the evidence already retrieved in the previous tool outputs and the context of the prior findings. Let's review the active threads and see if we can resolve or update anything: 1. `019ea66e-0e2e-79f7-bb6c-402131a94c7b` (Eightfold AI Motion to Dismiss Hearing on FCRA/ICRAA Claims, scheduled for August 4, 2026). Since today is July 27, 2026, the hearing has not occurred yet. So this watch is still active and has not fired. 2. `019f5ab0-d24e-706b-9680-d9be684f952d` (OpenAI Deadline to Respond to Florida AG Lawsuit, due August 24, 2026). Today is July 27, 2026, so this deadline has not occurred yet. This watch is still active and has not fired. 3. `019eca7f-c58b-7d75-98d1-cc254fa58385` (Florida AG v. OpenAI Motion to Dismiss or Procedural Ruling). Since the response deadline is August 24, 2026, no ruling has occurred yet. This watch is still active and has not fired. 4. `019eca7f-b87f-7990-bfc4-8c546e990794` (Illinois IDHR Re-introduction of AI Employment Notice Rules). Let's see if we have any info. The search budget is spent, so we couldn't find any new info this cycle. This watch remains active. 5. `019ea66e-1a44-7417-a11f-5a37bfc609ee` (EU Product Liability Directive Transposition Deadline, December 9, 2026). This is far in the future, so this watch is still active. Wait, since we don't have any new findings to write up beyond what we already have in our notes and what was retrieved, let's see if we should write or update any notes. Wait! Let's check if the user wanted any specific findings updated. The brief says: "Track how global regulators are approaching AI liability: new legislation and proposals across jurisdictions, enforcement actions, court decisions, regulatory guidance documents, industry compliance frameworks, and shifts in how liability is being assigned between developers and deployers. Surface emerging trends a legal or risk team at an enterprise need to stay current on." Let's look at the search results we got. We searched for Eightfold and got details about the complaint and its legal theories: - It alleges Eightfold is a Consumer Reporting Agency (CRA) under the FCRA and an investigative CRA under California's ICRAA. - It targets Eightfold's "Evaluation Tools" which evaluate and rank job applicants using data gathered from applications, employer internal data, external data, and Eightfold's proprietary LLM. - It claims that Eightfold's evaluation of "likelihood of success" on a 0-5 scale and candidate profiles constitute "consumer reports" under the FCRA. - Crucially, it highlights that once an applicant applies, Eightfold retains that data and uses it to evaluate other applicants or the same applicant for other positions, creating a continuous "dossier" without consent. - This is a novel legal weapon (FCRA/ICRAA) that bypasses the traditional employment discrimination framework (Title VII/ADEA), creating direct liability for AI vendors and potentially the employers who deploy them. Let's see if we can update the note `kistler-v-eightfold-ai-fcra-icraa-class-action-2026` or if we should write a new note. The existing note is titled "FCRA and ICRAA Class Action Against Eightfold AI: Briefing Completed Ahead of August 2026 Motion to Dismiss Hearing" (slug: `kistler-v-eightfold-ai-fcra-icraa-class-action-2026`). Let's update it to reflect the specific compliance and risk implications of the lawsuit's core allegations for enterprise risk teams, using the quotes we retrieved from the complaint and the Ogletree Deakins article. Let's carefully construct the updated note for `kistler-v-eightfold-ai-fcra-icraa-class-action-2026`. Wait, let's look at the quotes we have: From the Ogletree Deakins article: - "The Eightfold lawsuit could be the first of a new type of class action litigation targeting employment tools that use AI or automated or algorithmic decision-making processes that retrieve or use applicant- or employee-specific data." - "According to the complaint, Eightfold generates reports on applicants using “AI-powered tools that assemble and evaluate information on prospective employees” and assess their “suitability” for a job based on factors such as “work history, projected future career trajectory, culture fit, and other personal characteristics.”" - "Further, the complaint alleges that once an applicant applies for a job with an employer using the Eightfold tool, Eightfold retains that applicant’s data and uses it to evaluate other applicants for the same job, unrelated positions, or “for that same job applicant for other positions in the future.”" From the complaint (`Kistler et al. V eightfold AI Inc. Complaint`): - "Eightfold is a Consumer Reporting Agency (“CRA”) for FCRA purposes because it is in the business of assembling and evaluating information on consumers for the purpose of furnishing consumer reports (including information bearing on job applicants’ character, general reputation, personal characteristics, or mode of living, which is expected to be used as a factor in establishing the consumer’s eligibility for employment purposes) to third parties, including Microsoft, Paypal and many more." - "Eightfold is also an investigative CRA under California law because it is in the business of collecting, assembling, evaluating, compiling, reporting, transmitting, transferring, or communicating consumer information regarding consumers’ character, general reputation, personal characteristics, or mode of living (consumer reports), including for employment purposes, to companies, including prospective employers." Let's integrate these into `kistler-v-eightfold-ai-fcra-icraa-class-action-2026`. This is a great way to refine the existing finding by providing the exact legal mechanisms and quotes from the

I will now update the existing note `kistler-v-eightfold-ai-fcra-icraa-class-action-2026` with the concrete allegations and quotes retrieved from the complaint and legal analysis, highlighting the specific legal mechanisms and risk dimensions of the FCRA/ICRAA claims.

Now that the note has been successfully updated with the concrete quotes and legal mechanisms of the Eightfold AI case, I will finish the research cycle.

I have successfully completed the research cycle. ### Summary of Actions Taken: 1. **Reviewed Active Threads and Watches**: Examined the existing threads and watches on the topic, including upcoming litigation deadlines for *Kistler v. Eightfold AI Inc.* (August 4, 2026) and *Florida AG v. OpenAI* (August 24, 2026). 2. **Conducted Evidence Gathering**: Searched for details on the Eightfold AI litigation, retrieving specific allegations and direct quotes from the active complaint and legal analysis. 3. **Refined Existing Findings**: Updated the note `[[kistler-v-eightfold-ai-fcra-icraa-class-action-2026]]` with detailed legal arguments, including the plaintiffs' theory of AI platforms as Consumer Reporting Agencies (CRAs) under the FCRA and investigative CRAs under California's ICRAA, as well as the risk implications of continuous data retention and re-evaluation. 4. **Completed the Cycle**: Finalized the cycle using `finish_cycle`.