First American’s 2019 leak tested the route from a pen test to the 8-K

First American’s own testers had reported the EaglePro flaw by January 2019. The SEC and New York regulators later showed how the finding stalled before it reached the people who had to disclose it. Each of four handoffs failed separately.

Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.

Two-minute summaryThe dek and the first sentence of each section, quoted from the article

First American’s own testers had reported the EaglePro flaw by January 2019. The SEC and New York regulators later showed how the finding stalled before it reached the people who had to disclose it. Each of four handoffs failed separately.

  1. The week On the morning of May 28, 2019, which the SEC’s later order called the first trading day after the journalist’s article was published, First American Financial Corporation furnished a Form 8-K to the Securities and Exchange Commission.
  2. From the test report to the remediation queue The SEC order describes the first handoff in detail.
  3. From the application to the data inventory The second handoff concerns what the company believed the application held.
  4. The design of the link itself Before turning to the people, it is worth staying with the technology for a moment, because the flaw was of a kind any firm that shares documents through links can check for.
  5. From the security team to the executives The third handoff failed under pressure, in the days the journalist’s article forced a response.
  6. From the executives to investors, regulators and customers The fourth handoff is the one the public saw.
  7. What the board could have seen A board or audit committee does not read penetration test reports.
  8. The customer’s side of the exposure The documents in EaglePro were not First American’s own records in any simple sense.
  9. The strongest objection: a clerical error is not a governance failure The best defense of First American’s position runs like this.
  10. How the exposure differs by kind of firm The four handoffs exist in any financial firm, but the rules that apply to each differ.
  11. What a firm could have checked in June 2019 A financial services executive who read the First American reports in the weeks after May 24 could have asked four questions of their own firm, one for each handoff, without waiting for any regulator’s findings.
  12. With hindsight The regulators took their time.

The week

On the morning of May 28, 2019, which the SEC’s later order called the first trading day after the journalist’s article was published, First American Financial Corporation furnished a Form 8-K to the Securities and Exchange Commission. Attached was a press release headed “First American Financial comments on its ongoing investigation into reported information security incident.” The company said it had shut down external access to a production environment “with a reported design defect that created the potential for unauthorized access to customer data,” had retained an outside forensic firm, and that, though the investigation was in its early stages, there was “no indication that any large-scale unauthorized access to sensitive customer information occurred.” It said it would notify affected consumers and provide credit monitoring if the investigation showed confidential information had been compromised, and that it would post updates exclusively on a page of its website, where consumers who believed their information had been compromised would soon be able to report it.

The filing answered events of May 24. According to the SEC’s later order, on the morning of May 24, 2019, a cybersecurity journalist emailed First American’s investor relations staff that its web application was “leaking more than 800 million documents from real estate transactions going back from 2003 to the present day,” including bank account numbers, mortgage and tax records, Social Security numbers, wire transaction receipts and driver’s license images. After the close of trading that day the journalist published an article describing how changing the document number in a link to one customer’s file brought up other people’s records. The New York department’s consent order later gave the journalist’s figure as 885 million documents. First American gave the journalist a statement that it had learned of “a design defect in an application that made possible unauthorized access to customer data” and had shut down external access to the application.

The application was EaglePro, a proprietary system the company launched in 2013, which First American used to share images of title and escrow documents with parties to real estate transactions. According to the New York consent order, an October 2014 update let authorized employees select documents, after which EaglePro emailed the recipient a link to them. First American is a title insurance and settlement services company headquartered in Santa Ana, California. Its title insurance subsidiary is licensed by the New York State Department of Financial Services, which made it a covered entity under New York’s cybersecurity regulation, 23 NYCRR Part 500.

According to the department’s later consent order, First American notified it of the vulnerability on May 27, 2019, as Section 500.17(a) of that regulation requires. On May 31, 2019, in an “Incident Update” to customers, the company acknowledged that documents containing nonpublic personal information were potentially able to be accessed, and offered complimentary credit monitoring to anyone concerned who had received a title insurance policy or escrow or closing services from First American on or after January 1, 2003.

According to the SEC, the executives who approved the May 24 statement and the May 28 filing did not know that the company’s own security staff had found the same flaw months earlier. That fact came out later, through the regulators.

The First American exposure is remembered for its size: hundreds of millions of document images, reachable by anyone who changed a number in a link. The size is real, but it is not what made the case instructive for financial services firms. What made it instructive is that the company had found the flaw itself. A manual penetration test had described the problem in writing in January 2019. The report reached the people responsible for fixing it. Then the finding stalled, and when a journalist emailed in May, the people who had to speak to investors and regulators did not know the company had been warned.

Two regulators drew separate charges from the same facts. The SEC charged a failure of disclosure controls: the executives who approved the May 2019 statements lacked information the company held. The New York Department of Financial Services charged failures of access controls and of risk-based cybersecurity policy. Neither regulator alleged that the penetration test was badly done. Both cases turned on what happened to its result.

We read the case as a chain of four handoffs, each of which failed in a different way. The first ran from the test report to the remediation queue. The second ran from the application to the data classification that should have told the company how much was at stake. The third ran from the security function to the executives. The fourth ran from the executives to investors, regulators and customers. A firm can have a capable security team and still fail at each link, because each link belongs to a different part of the organization. Between the second and third handoffs we look at the design of the sharing link itself, since that set the size of the exposure. After the four handoffs we look at what a board could have seen, and at the exposure of the lenders and other parties whose customers’ documents were in the system. We then take the strongest objection to this reading, and set out how the exposure differs by kind of financial firm.

From the test report to the remediation queue

The SEC order describes the first handoff in detail. Between December 2018 and January 2019, First American’s information security staff ran a manual penetration test of EaglePro, an assessment of its internet-facing application. They finalized their report on January 11, 2019. It identified a vulnerability rated “serious,” or level 3, and described it as “access to PDF’s and OrderDetails without authentication.” Replacing the document identifier in the page address with another sequential number gave access to other, unrelated document sessions without authentication. The report also found that ordinary internet searches returned title documents viewable inside an EaglePro session, bypassing authentication.

The report was shared the same day with information security managers, the staff of the company’s vulnerability remediation management program, and the EaglePro “Accountable Remediation Owner,” the person the company’s policies made responsible for fixing vulnerabilities in that application. The New York consent order adds that preliminary findings had been shared with the remediation and application security teams on December 17, 2018, with the testers saying the finding should be “address[ed] as soon as possible.”

Then the record went wrong in a small way. Under the company’s policies, a level 3 vulnerability was medium risk and had to be fixed within 45 days. Through a clerical error, the vulnerability was entered in the automated tracking system as level 2, low risk, which carried a 90-day deadline. The tracking system calculated a remediation deadline of May 8, 2019. The policy said that if an owner could not remediate in time, the owner’s management had to contact information security about a plan, and if remediation was not technically possible or was too costly, the owner and management had to seek a waiver or risk acceptance from the chief information security officer. No waiver or risk acceptance was requested. The deadline passed. The flaw was still live on May 24.

Three things about this handoff are worth a CIO’s attention. First, the severity a finding carries into the tracking system controls everything downstream: the deadline, the escalation, and whether anyone senior sees it. A transcription error at intake changed the clock. Testers’ reports and the tracking record should be reconciled, so that a finding rated serious by the people who found it cannot enter the queue as low without someone noticing.

Second, the escalation path was written as a duty of the owner. When a deadline approached without a fix, the owner was supposed to raise it. If the owner did not, nothing in the account the SEC gives shows the system raising it instead. A missed deadline with no waiver on file is a fact a tracking system can detect. A design in which overdue findings without approved exceptions go automatically to the CISO does not depend on the person who is behind schedule volunteering the news.

Third, the test report itself said that the testers had found no nonpublic personal information in the documents they reviewed, but that “it is unknown if any additional documents” contained it, which “requires further investigation by the application owner.” The finding carried an open question about impact, and the question was assigned to the same person who owned the fix. The consent order states that no such further investigation or review by senior management was conducted before the journalist’s article. A finding whose impact is unknown should not be scored as if the impact were known to be low.

From the application to the data inventory

The second handoff concerns what the company believed the application held. The New York consent order says that as part of its risk assessment, First American classified EaglePro as an application that did not contain nonpublic information, even though the vulnerability could expose documents that did. Because of that classification, the department found, the company did not implement an appropriate, risk-based access control policy for EaglePro.

The SEC order explains how the classification could be wrong. EaglePro pulled images from a repository of about 800 million title and escrow document images, both public and nonpublic. Images with nonpublic personal information were supposed to carry an “SEC” legend. The tagging was done by hand, which “lent itself to a number of misclassifications.” A 2018 internal analysis had found that the repository potentially contained tens of millions of images with nonpublic personal information that were not tagged. Staff could send documents through “unsecure” EaglePro packages that required no password, and could be passed on to third parties. The consent order adds that EaglePro links had no expiration date until May 24, 2019, and that although users were told not to send nonpublic information through those links, no technical control stopped them.

So the company’s inventory said one thing and its data said another, and the company had evidence of the gap in its own 2018 analysis. Any application connected to that repository could reach the untagged documents. The risk of an application depended on what the repository held, not on what the application was meant to send.

For a financial firm, this is the handoff that does more than any other to decide the size of a breach. A vulnerability rated against an application’s intended data will be under-rated if the application can reach other data. The useful control is to rate applications by the highest sensitivity of data they can technically reach, not the data they are supposed to handle. When a firm knows its tagging is unreliable, as First American’s own analysis indicated, it should assume untagged records may be sensitive until shown otherwise.

It also shows why the penetration testers’ open question mattered. The testers did not find personal data in the documents they looked at, and said the rest needed investigation. The 2018 analysis suggested the answer. Had the two been connected, the finding’s impact would have been visible in January. The information existed in two places in the same company and was never put together.

Before turning to the people, it is worth staying with the technology for a moment, because the flaw was of a kind any firm that shares documents through links can check for. The SEC order describes it plainly. A user could take the address generated for an EaglePro package, which might link to an image containing personal information, and change the digits to view other document images the user should not have been able to see. Some images sent through unsecure packages had also been cached by public search engines.

Three design properties combined. The identifiers were sequential, so one valid link revealed the pattern for others. The application served a document to anyone presenting its address, without checking that the requester was entitled to that particular document. And links did not expire, so a document shared once stayed reachable indefinitely. Any one of these alone would have limited the damage. A random identifier would have made guessing impractical. A per-request entitlement check would have refused documents outside the recipient’s package. An expiry date would have closed old packages.

The search engine finding adds a fourth property. Pages reachable without authentication can be found and indexed by crawlers, after which the documents are visible to anyone searching, with no guessing needed. A firm that serves customer documents through links should assume those links may be posted, forwarded or crawled, and design so that a leaked link exposes one package for a limited time and nothing else.

These are checks a security team can run against any document-sharing feature in an afternoon: request a document with a modified identifier, request an old link, search for the application’s address pattern in public search engines. The EaglePro penetration test ran the first and the last of these and reported both. What the firm needed was not a better test, but for the result to change the application.

There is a forensic point as well. On May 28, the 8-K said there was no preliminary indication of large-scale unauthorized access to sensitive customer information. A statement like that can only be as good as the access records behind it. To know whether outsiders had walked the identifier sequence, the company needed logs of who requested which documents, from where, over the period the flaw was open, which the security staff later determined reached back to 2014. A firm that keeps access logs for a short period, or does not log document requests at all, cannot answer the question its regulators and customers will ask first. Log retention for customer-facing document systems should be set with that question in mind.

From the security team to the executives

The third handoff failed under pressure, in the days the journalist’s article forced a response. According to the SEC order, the company’s CISO learned about the January 2019 report, and that the vulnerability had not been fixed on time, for the first time on the morning of May 24, shortly after the journalist’s outreach. The CIO learned of it on May 25. Between May 24 and May 28, the CISO and CIO took part in numerous meetings with the senior executives responsible for the company’s disclosures, including the chief executive and the chief financial officer.

Yet, the order says, those executives were not made aware of the January report before the press statement on May 24 or the 8-K on May 28. They did not evaluate whether to disclose the company’s prior awareness of the flaw, or its failure to fix it. They did not know that the security staff had known for months. And after the 8-K was furnished, the security staff determined that the vulnerability had existed since 2014. The order’s finding with the widest reach for other firms is short: as of May 24, 2019, First American “did not have any disclosure controls and procedures related to cybersecurity, including incidents involving potential breaches” of its data.

Two lessons follow, and they point in different directions. The first is that a CISO who learns of the firm’s prior knowledge on the day of a crisis has no time to judge whether that fact matters for disclosure, and should not be the one deciding. The SEC order does not say why the information did not reach the executives in those meetings. It only says it did not. A disclosure process needs a standing question, asked of the security function at every incident, about what the company knew before the incident and when. The answer should go to the people approving disclosures in writing, whether or not anyone thinks it is material.

The second is that the CISO’s own late discovery was itself a finding. A serious vulnerability in an internet-facing application holding customer data had passed its deadline without the CISO knowing. That is the first handoff failing again, seen from the top. The disclosure control failure and the remediation failure were linked: the remediation process did not surface the finding to the CISO, so the CISO could not surface it to the executives until after a journalist had.

For firms reading the case in 2019, the practical step was to write down who must be told what during a cyber incident, by when, and through which channel, and to make the history of the affected system part of the package. The facts that matter for disclosure include the incident, the prior findings on the affected system, their status, and any exceptions granted. Those are records a well-run vulnerability program already keeps. The question is whether they reach the disclosure committee.

From the executives to investors, regulators and customers

The fourth handoff is the one the public saw. Between May 24 and May 31, First American spoke to four audiences through four channels. It gave a statement to the journalist and national media on May 24. It notified the New York department on May 27, as Part 500 required for a covered entity. It furnished an 8-K to the SEC with a press release on May 28. It posted an incident update for customers on May 31, acknowledging that documents with nonpublic personal information had potentially been accessible and offering credit monitoring to anyone who had received title insurance or escrow or closing services from the company since the start of 2003.

Each statement was tuned to its audience. The 8-K press release stressed that there was no preliminary indication of large-scale unauthorized access. The May 31 customer update acknowledged that documents containing nonpublic information had been exposed to potential access. Those two statements are consistent, since exposure is not the same as access, but they put weight on different facts. Neither mentioned the January penetration test.

A regulated financial firm may face a similar structure: several regulators with different notification rules, investors with their own disclosure regime if the firm is public, customers with state breach notification rights, and the press. The case shows the cost when those statements are drafted from incomplete information. The SEC did not charge First American with saying something false. It charged that the company lacked controls to make sure all relevant information was considered. In our reading, the same gap affected the statements to the department and to customers, although the SEC’s order addressed only the press statement and the 8-K.

A firm can reduce this risk with one factual record per incident, kept from the first hour, kept by a named owner, from which every external statement is drafted. The record should hold what is known, what is not known, what the company knew before the incident, and the sources for each. When the record changes, every audience that received a statement based on the earlier version should be considered for an update. The New York regulation’s notice requirement, the SEC’s disclosure regime and state notification laws each have their own triggers, so the firm’s counsel should decide what each requires. But they should all decide from the same facts.

Wording choices deserve the same discipline. Phrases such as “design defect,” “potential for unauthorized access” and “no indication of large-scale unauthorized access” are each defensible. Each also invites a follow-up question: how long the defect existed, whether the company knew of it, and what evidence supports the absence of access. A disclosure committee reviewing such wording should ask what the honest answer to each follow-up would be, and whether the firm can give it. If the answer to “did you know before?” would be yes, that fact belongs in front of the committee before the statement goes out, whatever the committee then decides to say. The SEC’s case against First American rested on exactly that gap.

What the board could have seen

A board or audit committee does not read penetration test reports. It reads summaries, and the summaries are built from the same tracking system that mis-keyed the EaglePro finding. That makes the first handoff a governance question as well as an operational one. If the measures a board receives are counts of open findings by severity, a serious finding entered as low risk disappears into the wrong row, and the board sees nothing amiss.

Measures that would have exposed the EaglePro problem are easy to describe. The first is the number of findings past their remediation deadline without an approved exception, by severity, with the oldest named. The second is the number of exceptions and risk acceptances granted, by whom, and for which systems. The third is the share of internet-facing applications whose data classification has been confirmed against what they can actually reach. The fourth is the time between a finding’s discovery and its arrival on the CISO’s desk when it is serious. Each can be computed without judgment. The first would have flagged EaglePro as soon as May 8, 2019, the deadline the tracking system itself had calculated, passed without a fix or an exception.

There is a second governance point. The SEC order describes the CISO learning of the January report on the morning a journalist called. A board that wants to know whether its CISO can see the firm’s own findings can ask the CISO to name, without preparation, the oldest open serious finding on an internet-facing system holding customer data. If the CISO cannot, the reporting design is the problem, not the CISO.

Boards also set the tone for exceptions. The EaglePro policy had a route to a waiver or risk acceptance, approved by the CISO, for vulnerabilities that could not be fixed in time. That route was not used. An exception process works only if asking for an exception is safer for the owner than letting a deadline lapse in silence. Where missed deadlines carry no consequence and exceptions draw scrutiny, owners have reason to say nothing. A board can ask for both numbers, missed deadlines and exceptions granted, side by side, and treat a low exception count paired with a high overdue count as a warning.

The customer’s side of the exposure

The documents in EaglePro were not First American’s own records in any simple sense. According to the journalist’s email quoted in the SEC order, they included bank account numbers, mortgage and tax records, Social Security numbers, wire transaction receipts and driver’s license images, and the New York order says the company collected documents from customers, real estate agents and other parties to transactions. The consent order describes First American collecting documents from those parties, some containing personal information, and storing them in its repository.

That makes the case a third-party risk case for every lender, broker and agent who sent borrower and client documents to a title or settlement company. For a mortgage lender, the settlement agent is a service provider that can receive sensitive files from a loan, as the document types in the regulators’ accounts show. The lender’s due diligence on that provider may have covered licensing, insurance, escrow controls and wire fraud procedures. The EaglePro facts suggest additional questions a lender could put to any provider that receives its customers’ documents: how the provider classifies what it receives, how it shares documents onward, whether shared links expire or can be guessed, how long it keeps documents, and how quickly it will tell the lender if they are exposed.

Retention deserves its own line. The journalist’s report, as both regulators recount it, said the exposed images dated back to 2003. The customer update offered credit monitoring to anyone who had received title insurance or escrow or closing services since then. The longer a provider keeps documents online and reachable, the larger the population that a single flaw can expose, and the further back a notification exercise has to reach. A lender cannot set its provider’s retention policy, but it can ask what it is, and whether documents from closed transactions remain reachable through the same applications as live ones.

Notification also runs through the customer relationship. When a provider is exposed, the people affected may contact the lender, not the provider, because the lender is the name they know. A lender that learns of a provider’s exposure from the news has to answer its own customers without the facts. Contract terms that require prompt notice to the lender, and a named contact at the provider for incidents, shorten that gap. We think lenders and other financial firms that send documents to settlement providers should treat those providers’ document-sharing systems as part of their own exposure, and ask for evidence accordingly.

The strongest objection: a clerical error is not a governance failure

The best defense of First American’s position runs like this. The company had a penetration testing program, and it worked: it found the flaw. It had a remediation policy with deadlines, escalation and a waiver process. The failure began with a clerical error in entering a severity rating, and was compounded by one owner not following policy. Every large organization has clerical errors. Regulators charged the company with systemic failures based on individual lapses.

We think this objection is strongest on the first handoff and weakest on the others. A mis-keyed severity is an individual error. But the program let a serious finding pass a deadline without anyone above the owner learning of it, and that is a design property. The application’s classification as holding no nonpublic information, in the face of the company’s own analysis that tagging was unreliable, was not a clerical slip. And the SEC’s finding that the company had no disclosure controls related to cybersecurity at all describes the absence of a process, not a lapse in one.

There is also an objection on the disclosure side. The 8-K said there was no preliminary indication of large-scale unauthorized access to sensitive data. Nothing in the SEC’s order says that statement was untrue when made. Knowing about the January report might not have changed it. That is fair, and the SEC charged only the narrower failure. But the narrower failure is the more general lesson. A disclosure that turns out to be accurate is still made without adequate controls if the people approving it lacked relevant facts the company held. Firms cannot rely on being lucky in what they did not know.

A third objection concerns hindsight. The flaw, the security staff later found, had existed since 2014, long before the January 2019 test. On that view, the test was the first time anyone could have known, and the firm should be judged only on the months between the report and the article. We agree that the months are what the regulators judged. But the 2014 date matters for a different reason. It shows that the company’s earlier security reviews, whatever they were, had not caught a flaw that a manual test later found. A firm that tests an internet-facing application holding customer data for the first time in years, or tests it only with automated scanners, should expect that the first manual test may find something old.

A fourth objection is about size. The penalties, $487,616 from the SEC and $1 million from the New York department, were small relative to a public company’s finances. A skeptical reader might conclude the regulatory cost of the failures was modest. We do not draw that conclusion. The penalties came after a regulatory process that, in New York, ran from a statement of charges in July 2020 to a consent order in November 2023, with amended charges in between. The record each regulator built became public. And the consent order bars the company from seeking reimbursement of the New York penalty through insurance.

How the exposure differs by kind of firm

The four handoffs exist in any financial firm, but the rules that apply to each differ.

Public companies. First American’s SEC exposure came from being an issuer with securities registered under the Exchange Act, subject to Rule 13a-15’s requirement to keep disclosure controls and procedures. Any public financial company, whether a bank holding company, an insurer, a broker-dealer parent or a fintech, carries the same obligation. Rule 13a-15 applies to issuers whatever their industry, though the order did note that First American’s business involved real estate transaction data. The charge rested on cyber incidents being information that might have to be disclosed, and on the company having no disclosure controls related to cybersecurity.

Firms licensed in New York. First American’s title insurance subsidiary was a covered entity under Part 500 because the department licensed it. The consent order describes Part 500 as placing obligations on all DFS-regulated entities, including: a cybersecurity program based on a risk assessment, written policies that address areas including data governance and classification, access controls and identity management, limits on user access to systems holding nonpublic information, and notice to the department of qualifying cybersecurity events. The department’s findings in this case rested on the policy and access control sections, Sections 500.3 and 500.7; the order records that the company gave the Section 500.17(a) notice as required.

Private firms and smaller institutions. A privately held firm outside New York would not have faced either of these charges in the same form. It would still have faced state breach notification laws, customer contracts and, for banks and credit unions, its own prudential regulators’ expectations. The four handoffs are still the places where a breach grows or shrinks. The difference is which regulator, if any, will later ask for the record.

Firms that share documents with outside parties. EaglePro existed to send documents to buyers, sellers, agents and lenders who were not First American’s employees. Mortgage lenders, wealth managers and others who share statements and closing documents through links face the same design choices: whether links expire, whether they need a password, whether they can be forwarded, and whether the identifiers in them can be guessed. The SEC order notes that unsecure packages required no password from the recipient and could be shared onward with third parties, who needed no password either. A sharing feature built for convenience becomes the attack surface when its identifiers are sequential. Firms that offer such features to customers, agents or counterparties should include them in penetration testing scope by name, and should treat findings against them as findings against the underlying document store, since that is what an attacker reaches.

What a firm could have checked in June 2019

A financial services executive who read the First American reports in the weeks after May 24 could have asked four questions of their own firm, one for each handoff, without waiting for any regulator’s findings.

  • Can the vulnerability tracking system show every finding past its deadline without an approved exception, and does that list go to the CISO without anyone choosing to send it?
  • Are applications rated by the highest sensitivity of data they can reach, including data that may be mislabeled?
  • When an incident occurs, is there a written step that asks what the firm knew about the affected system beforehand and sends the answer to those approving disclosures?
  • Does one incident record feed every statement to regulators, investors and customers?

None of those questions needs new technology. Each needs an owner and a record. First American had a testing program, a tracking system and policies. What it lacked, on the regulators’ accounts, was the connection between them.

With hindsight

The regulators took their time. The New York department filed a statement of charges and notice of hearing on July 21, 2020, seeking a civil penalty and an order requiring First American to remedy the violations, and amended it on March 10, 2021 and June 17, 2021. The SEC settled first. Its order, dated June 14, 2021 and announced the next day, found that First American had violated Exchange Act Rule 13a-15(a) by failing to maintain disclosure controls and procedures designed to ensure that all available, relevant information about the vulnerability was analyzed for disclosure. Without admitting or denying the findings, the company agreed to a cease-and-desist order and a $487,616 penalty. The New York case ended with a consent order approved by the acting superintendent on November 27, 2023. The department found violations of Section 500.3, for failing to maintain and implement an effective cybersecurity policy for access controls based on its risk assessment, and Section 500.7, for failing to implement access controls sufficient to prevent unauthorized users from gaining access to nonpublic information through EaglePro. The company agreed to pay a $1 million penalty and not to seek reimbursement through insurance.

Two years later, the SEC wrote disclosure expectations of this kind into a rule; the adopting release does not cite the First American order. On August 4, 2023, the SEC published its final rule on cybersecurity risk management, strategy, governance and incident disclosure, effective September 5, 2023. It added Item 1.05 to Form 8-K, requiring a report within four business days after a registrant determines that it has experienced a material cybersecurity incident, with compliance from December 18, 2023 for registrants other than smaller reporting companies. It added Item 106 to Regulation S-K, requiring annual disclosure of a registrant’s processes for assessing, identifying and managing material cybersecurity risks, management’s role, and the board’s oversight. The rule said that adhering to normal internal practices and disclosure controls and procedures would suffice to demonstrate good faith compliance with the timing of the materiality determination. What the 2021 order punished as an absence of process, the 2023 rule made a disclosed process.

First American later used the new item itself. On December 22, 2023, it filed an 8-K under Item 1.05 reporting unauthorized activity on certain of its information technology systems, and saying it had isolated systems from the internet on December 20, could not estimate the duration or extent of the disruption, and could not yet determine whether the incident would have a material impact. On January 12, 2024, an amended report said the company believed it had contained the incident and resumed operations, that some fourth-quarter transactions had been delayed and others moved to other providers, and that it expected a material impact on fourth-quarter 2023 results of operations but not on its overall financial condition.

That sequence shows how the disclosure regime changed around the same company. The 2019 incident was disclosed under a general disclosure regime by executives who lacked key facts. The 2023 incident was disclosed under a specific item, with its uncertainty stated and a later update. As of November 14, 2026, the eCFR’s version history shows Regulation S-K Item 106 unchanged since it took effect on September 5, 2023.

Spotted an error? Tell us and we will check it and correct the article where needed.

  • First American
  • NYDFS Part 500
  • SEC disclosure
  • vulnerability management
  • data classification