People are increasingly using AI tools to ask questions they might previously have entered into an internet search engine or discussed privately with a friend. In a contentious custody case, however, a search such as “how can I keep my children from my spouse?” could potentially take on a very different significance when viewed in the context of litigation.

Family law attorneys may begin seeking AI prompts, conversation histories, and other records during discovery or through subpoenas, particularly when those records could shed light on a party’s intentions, state of mind, or conduct. A series of searches concerning hiding assets, restricting parenting time, monitoring a spouse, or keeping children away from the other parent could become relevant evidence depending on the facts of the case and applicable discovery rules.

Although traditional rules of discovery and evidence still apply, the use of AI presents a new and evolving area of uncertainty. Courts and practitioners may have to grapple with questions about the relevance, discoverability, authenticity, privacy, and admissibility of AI-generated conversations and search histories. Adding to the uncertainty, whether an AI provider retains the requested information, and whether it can or will disclose that information in response to a subpoena or other legal process, may vary significantly depending on the platform, its policies, and the circumstances of the request.

Nevertheless, the practical lesson for family law litigants is important: AI searches should not necessarily be treated as consequence-free conversations. Just as texts, emails, social-media activity, and internet searches can become relevant in litigation, AI interactions may increasingly become another source of electronically stored information.

For attorneys, this creates a new area of discovery to consider. For clients, it creates a new reason to think carefully about what is entered into an AI platform, particularly when a divorce, custody dispute, or other family law proceeding is anticipated.

As AI becomes more integrated into daily life, family law discovery will likely have to evolve with it.

Electronic surveillance may take many forms. Allegations can involve unauthorized access to emails or social media accounts, monitoring text messages, tracking a person's location through a device or vehicle, or the use of technology to monitor communications. In some cases, a party may not realize that their activity is being monitored until evidence of the surveillance emerges during a divorce or custody proceeding.

These allegations can have significant legal consequences. Depending on the circumstances and applicable law, information obtained through unauthorized surveillance may raise questions about how the evidence was obtained and whether it may be used in court. Surveillance allegations may also become relevant to broader issues in a custody case, particularly when the conduct suggests an effort to intimidate, control, or interfere with the other parent's privacy or safety.

Electronic surveillance can also create discovery issues. A party may seek records concerning devices, accounts, applications, location data, or other electronically stored information. At the same time, attorneys must carefully consider the privacy implications and the legal limits concerning the collection and use of that information.

The important lesson is that technology is an inherent part of family law, not a separate consideration. As electronic devices and digital accounts become increasingly integrated into everyday life, disputes over monitoring and surveillance are likely to become more common. Anyone who believes they are being electronically monitored, or who is accused of engaging in surveillance, should seek legal advice before accessing, copying, deleting, or distributing potentially relevant digital information.

Expanded Definition of “Personnel Record,” Record Retention Requirement

The new law defines "personnel record" as records that are used, have been used, or may be used in connection with decisions concerning an employee's qualifications for employment, promotion, transfer, compensation, or discipline. The definition also extends to records maintained by third-party vendors acting on behalf of an employer.

As a result, employers may need to evaluate not only the contents of formal personnel files, but also records maintained through payroll providers, HR platforms, and other personnel administration services. Employers will also have to determine whether manager working notes, internal communications regarding employee performance, draft evaluations, or portions of investigative files are subject to disclosure. Although the law excludes certain information that would constitute an “unwarranted invasion” of another individual's privacy, it provides limited guidance on how that exclusion should operate in practice when employers are balancing employee access rights against confidentiality concerns.

Employers must retain complete personnel records during an employee's employment and for three years following separation.

When Access and Notice Requirements are Triggered

The statute has a straightforward requirement for scenarios where an employee requests access to their file. Upon receiving a written request, employers will be required to furnish personnel records within five business days at no cost to the employee. Employees are entitled to review their personnel records twice in a calendar year.

But the statute creates an unclear requirement for scenarios where an employer adds items into the personnel file that could “negatively affect” the employee. Specifically, if there is any item added to the personnel file that could “negatively affect the employee's qualification for employment, promotion, transfer, additional compensation or the possibility that the employee will be subject to disciplinary action,” it must notify employees within 10 days.

In its current form, the law does not define what constitutes “negative” information, which will likely create substantial compliance questions. For example, it is unclear whether routine coaching memoranda, attendance records, customer complaints, investigatory materials, manager notes, performance improvement plans, or preliminary performance concerns trigger the notice obligation. Similarly, employers may struggle to determine whether information becomes "negative" only after a disciplinary decision has been made or whether notice is required whenever a document could potentially influence a future employment decision.

Right to Respond to Information in File

The new law allows employees to submit written statements contesting information contained in their personnel records. If the employer and employee cannot agree on whether the challenged information should be removed from the personnel file, the employer will be required to include the employee’s contention in their personnel file. The statute also provides employees with an avenue to expunge contested information through the “judicial process” with relatively little guidance regarding how disputes involving subjective performance evaluations or disputed factual conclusions should be resolved.

Government Entity in Charge of Enforcing Statute

The New York State Attorney General’s Office is responsible for enforcing this new law. Violations of the statute may result in civil penalties ranging from $500 to $2,500. The law also prohibits retaliation against employees who exercise their rights under the statute.

As a result of the ambiguities surrounding what information constitutes a “personnel record,” when information added to a personnel file “negatively affects” an employee, and the particular process an employee needs to follow in order to successfully “expunge” an item in their personnel record, the New York State Attorney General’s Office may weigh in to provide guidance on how it plans to enforce this law. In the meantime, employers will be required to make judgment calls regarding the scope of the notice requirement until courts or regulators provide guidance.

Conclusion

With the November 8, 2026, effective date rapidly approaching, employers should begin reviewing personnel record practices, assessing the records maintained by third-party vendors, updating retention protocols, and developing procedures for responding to employee requests within the statute's five-business-day deadline. Employers should also consider training human resources personnel and managers regarding the law's notice requirements, particularly given the uncertainty surrounding what information may be considered "negative."

While the law clearly reflects New York's intent to expand employee access to personnel information, many of its most consequential provisions remain open to interpretation. As employers prepare for implementation, the greatest compliance challenge may not be producing personnel records, but determining when the statute requires notice of information that could be viewed as having a negative impact on an employee's career. Until additional guidance emerges, employers should take a cautious approach and closely monitor developments surrounding this significant change to New York employment law.

The Proposed Arrangement

The Requestor, an orthopedic surgery provider, offers patients an optional concierge program that includes post-operative recovery support services and products for one (1) year following surgery. These services include wellness coaching, educational support, nutritional programs, digital health monitoring tools, and recovery-related products. The provider certified that none of these concierge services are reimbursable by Medicare, Medicaid, or commercial insurance, although the underlying surgical procedures may be covered by Federal health care programs.

Patients who elect to participate pay a separate concierge fee and execute a membership agreement before surgery. Under the proposed arrangement, the provider would warrant that a patient's substantial compliance with the concierge program would result in the patient not requiring revision surgery within two (2) years of the initial procedure. If the patient nevertheless required revision surgery during that period, the provider would refund the concierge fees paid in connection with the original surgery. The provider would not refund medical expenses, cover revision surgery costs, or provide any additional remuneration. Nor would the patient be required to return to the provider for the revision procedure.

Anti-Kickback Statute Analysis

OIG recognized that the arrangement potentially implicates the AKS because the offer of a refund could make the provider more attractive to prospective patients seeking surgeries that are reimbursable by Federal health care programs. As a result, OIG examined whether the arrangement satisfied the regulatory safe harbor for warranties found at 42 C.F.R. § 1001.952(g).

The warranty safe harbor protects certain written undertakings to refund, repair, replace, or provide other remedial action when an item, bundle of items, or related services fail to meet specified performance standards. OIG concluded that the proposed concierge fee refund met the regulatory definition of a warranty because:

• The commitment was memorialized in a written membership agreement
• The warranty formed part of the bargain between the provider and the patient
• The provider agreed to take remedial action in the form of a refund
• The refund would be triggered by the failure of the concierge program and related services to achieve the promised outcome of avoiding revision surgery within two years

OIG further determined that the arrangement satisfied the applicable conditions of the warranty safe harbor. Among other things, the provider certified that it would accurately disclose and document any refund, require patients to provide information to government authorities upon request, and refrain from conditioning the warranty on exclusive use of the provider or minimum purchase requirements.

Because the arrangement fit squarely within the warranty safe harbor, OIG concluded that it would not constitute prohibited remuneration under the AKS.

Beneficiary Inducements Civil Monetary Penalties Analysis

OIG likewise found no violation of the Beneficiary Inducements Civil Monetary Penalties (CMP). Although the refund offer could arguably influence a Medicare or Medicaid beneficiary's selection of a provider, the CMP's definition of remuneration excludes practices that are permissible under an AKS safe harbor. Having determined that the proposed arrangement qualified for protection under the warranty safe harbor, OIG concluded that the arrangement likewise posed no risk under the Beneficiary Inducements CMP.

Key Compliance Considerations

Several facts were central to OIG's favorable determination:

• The concierge services were separately purchased at fair market value
• The refund applied only to the concierge fees paid by the patient and not to medical, hospital, or surgical expenses
• Patients were not required to use the provider for any future procedure, including revision surgery
• The provider did not condition the warranty on exclusive use arrangements or minimum purchasing commitments
• Appropriate reporting, documentation, and disclosure requirements were incorporated into the membership agreement

Conclusions

Advisory Opinion 26-12 demonstrates OIG's willingness to recognize properly structured outcome-based warranty programs tied to non-covered services that satisfy the warranty safe harbor. The opinion may be particularly relevant to providers developing concierge, care coordination, recovery support, or other value-added patient programs.

At the same time, OIG emphasized that its approval was limited to the specific facts presented. Notably, the agency distinguished the proposed arrangement from the provision of free concierge services or other free benefits to patients. OIG reiterated its longstanding concerns that free items or services connected to federally reimbursable care may present significant fraud and abuse risks and would not necessarily qualify for safe harbor protection.

While advisory opinions bind only the requesting party, Advisory Opinion 26-12 offers a useful roadmap for providers seeking to design patient-centered warranty programs that reward outcomes without running afoul of the AKS or Beneficiary Inducements CMP. Proper structuring, fair market value pricing, careful documentation, and strict adherence to the warranty safe harbor remain essential to achieving a favorable compliance result.

While these efficiencies can accelerate the M&A process, speed should not be mistaken for judgment. An acquisition is not just about collecting and summarizing information. Buyers must determine what that information means for the value of the business, the structure of the transaction, and their willingness to proceed. AI can be a very helpful tool, but it cannot replace the wisdom and judgment that comes with experienced M&A counsel.

The better question is not whether AI can buy a company. It is how buyers and their advisors can effectively use AI tools without losing sight of the human judgment that ultimately protects the deal.

AI Can Review Documents, But It Cannot Assess the Buyer’s Risk 

During due diligence, AI tools can be very useful in taking over time consuming tasks such as:

• Summarizing contracts
• Locating change-of-control or assignment provisions
• Comparing similar agreements
• Flagging unusual or inconsistent terms
• Organizing documents by subject matter
• Identifying potentially missing information

AI’s speed here is particularly helpful in transactions involving hundreds or thousands of contracts. That said, identifying a provision is only the beginning of the analysis.

Suppose an AI tool finds change-of-control language in several agreements. The real questions are: Which agreements are critical to the business? Will consent be required before closing? Could the counterparty terminate or renegotiate? How would the loss of that relationship affect revenue, operations, or valuation?

The same language can create very different levels of risk depending on whether it appears in a minor vendor agreement or a contract with the target’s largest customer. AI may locate the provision, but it cannot reliably determine how much risk it creates for this particular buyer in this particular transaction. That requires a human’s understanding of the buyer’s objectives, the target’s business, the industry, and the broader structure of the deal.

AI Finds Issues, But Lawyers Determine Materiality

Due diligence often produces a long list of potential concerns, including incomplete employment agreements, gaps in intellectual property ownership, inconsistent customer contracts, regulatory deficiencies, or unresolved disputes. AI can help identify these issues, but it can also identify so many potential concerns that the buyer ends up with more information than clarity.

This is where experienced deal counsel can help the buyer separate meaningful risks from background noise. Not every issue warrants the same response, and many times, matters can be corrected before closing. Others may justify a purchase-price adjustment, escrow, or holdback. Certain risks may require a special indemnity, a closing condition or a change to the transaction structure. A sufficiently serious issue may cause the buyer to reconsider the deal altogether. But that kind of analysis depends on context, and determining how the pieces fit together remains a judgment-intensive exercise.

AI Does Not Negotiate the Deal

Every acquisition involves compromise. Even when the parties agree on price, they must negotiate representations and warranties, indemnification obligations, liability caps, escrows, closing conditions, earnouts, and numerous other provisions that allocate risk between the buyer and seller.

AI can propose language based on prior agreements or common market formulations. It may also help compare drafts and identify changes. But it is critical to note that it cannot be entrusted to manage the negotiation itself. That requires an understanding of what the other side wants, where the buyer has leverage, and which issues are worth pressing. It also requires the ability to recognize when a proposed solution creates a new problem elsewhere in the agreement.

Sometimes the strongest response is to hold firm. Other times, the better strategy is to address the concern through a price adjustment, special indemnity, post-closing covenant or alternative structure. A good deal lawyer does more than argue over language. They help the parties find a workable path to closing while protecting the client’s most important interests.

There is also a strong relationship component to every transaction. Many buyers need the seller, management team, or key employees to remain involved after closing, but an unnecessarily aggressive negotiation can damage the working relationship before the buyer takes control of the company. AI cannot read the room, understand personalities, or know when winning a drafting point could hurt the larger transaction.

AI Cannot Identify Every Question the Buyer Should Be Asking

AI is generally most effective when it has the right documents and receives the right inputs and instructions. The difficulty is that buyers do not always know what is missing, or even what they should be asking in the first place. Experienced M&A counsel can recognize patterns from prior transactions. They know which diligence requests are likely to uncover problems, which industries present specialized risks, and which seemingly routine answers require follow-up.

A contract summary will not help if a significant agreement was never uploaded to the data room. A review of the target’s intellectual property schedule may not reveal that a former contractor created critical software without signing an assignment. Employment records may look complete until someone asks how workers are classified, compensated, or managed in practice.

AI analyzes only the information it receives; whereas deal counsel knows when the available information does not tell the entire story. This ability to identify the unknown is particularly important for buyers entering a new industry or making their first acquisition. An experienced attorney can anticipate where issues tend to arise and tailor the diligence process to the buyer, the target, and the transaction.

AI Outputs Still Need to Be Verified

AI-generated summaries can sound authoritative even when they are incomplete or incorrect. A tool may overlook an exception, misinterpret a defined term, miss the interaction between multiple provisions, or reach a legal conclusion that the underlying language does not support.

In an M&A transaction, even a small error can have significant consequences. A missed consent requirement could delay closing. An incomplete summary of a customer agreement could distort the buyer’s view of recurring revenue. A mistaken interpretation of an indemnification provision could leave the buyer with less protection than expected.

AI should be treated as an analytical tool, not as the final reviewer or decision-maker. Its work must be checked against the underlying documents and evaluated by professionals who understand the legal and business consequences.

Buyers and their counsel must also consider how sensitive deal information is handled. Confidential financial records, customer data, employee information, and trade secrets should not be entered into an AI platform without understanding the tool’s security, retention, and privacy practices.

Using AI the Right Way in an Acquisition

The strongest acquisition teams will not ignore AI, but importantly, they will not treat it as a substitute for experienced advisors. AI is most valuable when it handles repetitive, time-intensive tasks and allows deal counsel to spend more time on higher-value work.

Buyers should have a firm understanding of:

• Which AI tools their advisors are using
• How confidential information is protected
• Whether AI-generated findings are independently verified
• How identified issues are prioritized and escalated
• Which decisions remain subject to human legal and business judgment

The goal in using these tools to help deal teams ask better questions, find issues sooner, and make better-informed decisions. AI may help a buyer move through diligence more efficiently, but it cannot decide whether a risk is acceptable, negotiate the right protection, or determine whether the deal still makes strategic sense. Those decisions will always require the experience, context, and judgment legal counsel and advisors bring to the table.

In the Celsius litigation, the court had to determine:

“1. Is the Litigation Administrator entitled to recover (a) the allegedly transferred digital assets if they remain in the applicable Defendant’s possession, custody or control or, alternatively, (b) their value?

2. If the Litigation Administrator is entitled to recover the value of any allegedly transferred digital assets, what value is the Litigation Administrator entitled to recover if the allegedly transferred digital assets:

i. Have appreciated since they were transferred from Celsius?

ii. Have depreciated since they were transferred from Celsius?”

In re Celsius Network, LLC, No. 22-10964-MG, 2026 WL 1999187, at *1 (Bankr. S.D.N.Y. July 10, 2026)

The litigation administrator claimed that the appropriate measure of damages is either (1) the return of the digital asset or (2) the market price of the asset if the asset appreciated. In the event the transferred asset has depreciated, the litigation administrator claimed that the court should award the market price of the digital asset on the date of the transfer.

Section 550 of the Bankruptcy Code allows a trustee or estate representative to recover either the property transferred or, if the court so orders, the value of that property. The Bankruptcy Code, however, does not specify how value should be measured when the property fluctuates in price after the transfer. Judge Glenn therefore turned to the statute's underlying purpose: restore the estate to the position it would have occupied had the transfer never occurred.

The Court's Three-Part Framework

Judge Glenn ultimately adopted a practical framework designed to balance three competing concerns:

• Making the estate whole
• Preventing preference defendants from profiting from avoidable transfers
• Avoiding unfair or potentially limitless liability resulting from cryptocurrency market volatility

The court held:

• For depreciating assets, the estate may recover the transfer-date value, regardless of whether the defendant still holds the asset.
• For appreciating assets still held by the defendant, the estate may recover the asset itself.
• For appreciating assets that have been sold, the estate may recover the sale price obtained by the defendant. The defendant bears the burden of proving both that the asset was sold and the amount for which it was sold. If the defendant cannot establish those facts, the estate may recover the judgment-date value.

This framework represents an effort to tailor traditional avoidance principles to the realities of cryptocurrency markets.

The Decision Regarding Depreciating Assets is Based on Well-Established Precedent

Judge Glenn's analysis began with what he viewed as the easier question: assets that declined in value after transfer. The court emphasized that the purpose of § 550 is restorative rather than punitive. If a cryptocurrency worth $100,000 at the time of transfer later falls to $10,000, requiring the estate to accept only the depreciated asset or its current value would leave the estate substantially worse off than if the transfer had never occurred. In effect, the bankruptcy estate would bear the entire market loss.

Relying on prior fraudulent transfer and avoidance precedent, Judge Glenn concluded that transfer-date value is the proper measure for depreciating property because it restores the estate to the financial position it would have occupied absent the transfer. The court reasoned that allowing only recovery of the current value would undermine the statute's remedial purpose.

Why the Court Rejected Judgment-Date Value for Appreciating Crypto

The more difficult issue involved assets that appreciated after transfer. The litigation administrator argued that appreciation should inure to the benefit of the estate because, had the transfer never occurred, the estate would have retained the asset and potentially realized the upside. The argument found support in cases involving real estate and other appreciating property. Judge Glenn, however, distinguished digital assets from many traditional forms of property. He noted that cryptocurrency is extraordinarily liquid, highly volatile, and can be sold almost instantaneously. Moreover, unlike real estate, many transferees no longer possess the specific assets that were transferred.

The court was particularly concerned that imposing judgment-date valuation on a defendant who sold cryptocurrency years earlier could create what it described as "essentially limitless liability." A customer who withdrew and sold Bitcoin long ago could be exposed to damages tied to market increases occurring long after the asset had been disposed of. The court concluded that such a result would be inequitable.

Judge Glenn also observed that there was no evidence that Celsius would necessarily have held the assets through the period of appreciation. The company may have sold, rebalanced, hedged, or otherwise deployed the assets. Awarding judgment-date appreciation therefore risked providing the estate with a windfall rather than merely restoring it.

The Court's Middle Ground

Rather than adopting either transfer-date value or full judgment-date value as a universal rule, Judge Glenn crafted a middle-ground approach.

If a defendant still possesses an appreciating digital asset, the estate may recover the asset itself and thereby receive the benefit of appreciation. That result mirrors what would have occurred had the transfer never happened. If the defendant sold the asset, however, recovery is limited to the benefit actually realized through the sale. The estate receives the sale proceeds, including any appreciation captured by the defendant, but not speculative gains that accrued after disposition.

In the court's view, this approach honors Congressional concern regarding a "wait-and-see" strategy by transferees while avoiding exposure to unlimited liability based solely on subsequent market movements.

Practical Takeaways for Digital Asset Holders

Judge Glenn's decision represents one of the first comprehensive attempts to address how cryptocurrency should be valued in bankruptcy avoidance actions. Rather than applying a rigid valuation date, the court adopted a flexible framework that distinguishes between appreciating and depreciating assets and considers whether the cryptocurrency remains in the transferee's possession. The ruling seeks to restore the estate and avoid imposing limitless liability driven solely by crypto market volatility. For digital asset holders, the message is clear: maintain thorough records, understand the risks of later-avoidable transfers, and recognize that post-transfer appreciation may not always belong to the party currently holding the coins.

Texas is a lot of things, but subtle is rarely one of them.

And there is certainly nothing subtle about Texas's recent efforts to attract businesses and capital to the Lone Star State. The latest evidence can be found in a term that has increasingly entered corporate nomenclature: DExit.

Short for "Delaware Exit," DExit refers to the growing trend of companies reconsidering Delaware as their state of incorporation. Much of the public discussion surrounding DExit has focused on a handful of high-profile corporate relocations and reincorporations. That focus, however, risks missing the larger story.

For more than a century, Delaware has occupied a singular place in corporate America. That dominance is remarkable when one considers Delaware's size. The entire State of Delaware could fit within the greater Houston metropolitan area and still leave room for a few Buc-ee's locations. Yet this geographically tiny state became the legal home of many of America's largest corporations by offering something few jurisdictions could match: a sophisticated body of corporate law and a specialized judiciary capable of resolving corporate disputes with predictability.

Texas has spent the last several years studying that playbook. Recent amendments to the Texas Business Organizations Code, the creation of the Texas Business Court and Fifteenth Court of Appeals, and the launch of the Texas Stock Exchange suggest that Texas is no longer content merely to attract a company's regional office or even its corporate headquarters. Texas now wants the company's legal domicile as well.

Whether Texas ultimately succeeds remains to be seen. Delaware's advantages remain substantial. But for the first time in decades, a serious challenger appears to be emerging.

Rewriting the Corporate Rulebook

The clearest evidence of Texas's ambitions appear in Senate Bill 29 and the legislature's recent amendments to the Texas Business Organizations Code (TBOC).

The legislature was not particularly coy about what it hoped to accomplish. State Senator Bryan Hughes described Delaware as a jurisdiction in which many companies had become "shackled by a burdensome Delaware legal establishment dominated by activist judges and special interest groups" and stated that the legislation would help bring American enterprise and jobs to Texas.¹ The bill's legislative history described Texas's goal as becoming the "corporate law capital of America."²

Amended TBOC § 21.218 narrows shareholder inspection rights by requiring a shareholder to hold shares for at least six months or own at least 5% of the corporation's outstanding shares before demanding access to books and records. The statute also limits inspections to requests related to a shareholder's economic interest in the corporation and generally restricts access to certain categories of electronic communications. That restriction on electronic communications is particularly significant. Emails, text messages, and other electronic communications are among the most voluminous and expensive categories of information for a company to collect, review, and produce upon a books and records request.

The legislature also codified the business judgment rule through TBOC § 21.419, establishing a statutory presumption that directors and officers acted in good faith, on an informed basis, and in the best interests of the corporation. Senator Hughes stated that the change would allow Texas businesses to "confidently deploy capital" by providing greater certainty to corporate decision-makers.³ 

Finally, new TBOC § 21.373 permits qualifying corporations to adopt heightened requirements for shareholder proposals, including minimum ownership thresholds, holding periods, and proxy solicitation requirements.

Reasonable minds may disagree on the wisdom of these changes, but their purpose is clear. Texas is actively reshaping its corporate-governance framework to make itself more attractive to corporate managers and directors considering where to incorporate.

Building a Texas Version of the Court of Chancery

Corporate lawyers have never chosen Delaware solely because of its statutes.

Delaware's true advantage has long been its Court of Chancery and the extensive body of precedent developed through decades of specialized corporate litigation.

Texas’s response is the Texas Business Court.

Operational since September 2024, the court has jurisdiction over certain complex business disputes, including corporate-governance matters and significant commercial transactions. For many claims, Business Court jurisdiction generally requires an amount in controversy exceeding $5 million. Appeals proceed directly to the newly established Fifteenth Court of Appeals, creating a centralized path for the development of Texas business law.⁴

Early results suggest the court is attracting substantial use. During its first year, the Texas Business Court received 185 filings, including 145 corporate-governance cases. Judges issued more than 680 orders, conducted more than 270 hearings and conferences, and produced 42 written opinions.⁵

Those numbers matter. Delaware's dominance did not emerge overnight. It developed because companies, lawyers, and judges repeatedly chose a singular and specialized forum for business disputes, which resulted in a predictable body of case law.

Texas appears to be attempting a similar process.

Ringing the Opening Bell

¹ Press Release, Office of Senator Bryan Hughes, Senator Bryan Hughes Files Groundbreaking Bill to Transform Texas Corporate Law (Feb. 27, 2025).

² S.B. 29, 89th Leg., Reg. Sess. (Tex. 2025), legislative history and bill analyses; Tex. Bus. Orgs. Code §§ 21.218, 21.373, 21.419.

³ Note 1, supra.

⁴ H.B. 19, 88th Leg., Reg. Sess. (Tex. 2023), legislative history; S.B. 1045, 88th Leg., Reg. Sess. (Tex. 2023), legislative history; Tex. Gov't Code ch. 25A and § 22.2151.

⁵ Office of Court Administration, The Business Court of Texas, Annual Report FY 2025.

⁶ Office of the Governor, Governor Abbott Marks Successful Trading Launch of Texas Stock Exchange (July 31, 2026).

⁷ Texas Stock Exchange, Texas Stock Exchange Celebrates Successful Launch of Trading (July 31, 2026).

⁸ Reuters, Texas Stock Exchange Lands First Primary Listings in Bid to Carve Out Market Turf (Aug. 18, 2026; updated Aug. 19, 2026).

⁹ Eric Revell, Texas Stock Exchange Officially Goes Live to Rival NYSE and Nasdaq, Fox Business (July 31, 2026).

¹⁰ Fortune Media, Amazon Claims No. 1 Spot on the Fortune 500 (June 3, 2026); Office of the Governor, Texas Leads With Most Fortune 500 Headquarters (June 3, 2026).

The headline for clients: No changes have been made yet, but many of the precedents adopted by the NLRB during the Biden administration may change over the next months or possibly years while the Trump administration remains in office.

A General Counsel memo does not change what the National Labor Relations Board has held is unlawful. It's a statement of prosecutorial priorities and the legal positions the GC's office intends to argue in pending and future cases. But GC memos are a reliable early-warning system for where Board law is going, and this one specifies, case by case, which Biden-era Board precedents the current GC is actively trying to unwind. The list includes essentially all of the new or revised interpretations of the National Labor Relations Act issued during the Biden administration.

If your organization has non-union operations, unionized operations, or is navigating an organizing campaign, several of these items are worth putting on your radar now.

GC Carey opens by reporting that the agency has resolved 9,247 pending cases since she took office, more than a 50% reduction in the backlog she inherited. Significantly, the regional offices are not required to route cases involving these targeted issues through the Division of Advice. Instead, they will keep investigating and prosecuting under existing Board law while the GC pursues these arguments through litigation.

Positions Already Being Argued in Pending Cases

These are precedents the GC's office has already asked the Board or an administrative law judge to overturn in specific, named cases:

Severance agreements and confidentiality/non-disparagement clauses. 
In Valley Radiology, P.A., the GC is arguing to overrule McLaren Macomb (2023) — the decision that made broad confidentiality and non-disparagement provisions in severance agreements presumptively unlawful. If the Board agrees, employers will regain more latitude to include standard confidentiality and non-disparagement language in severance and separation agreements without automatically committing an unfair labor practice.

Consent orders.
In the Amazon cases, the GC is asking the Board to overturn Metro Health/Hospital Metropolitano Rio Piedras (2024), which limited the Board's discretion to approve consent orders (settlement mechanisms) over the General Counsel's objection. A reversal would restore more flexibility for administrative law judges to approve settlements even without GC sign-off.

Work rules and handbook policies.
In Honeywell International, the GC is arguing to overturn Stericycle (2023), the standard that made facially neutral work rules unlawful if they could "chill" protected activity from the perspective of an economically dependent employee reading them in the worst reasonable light. A rollback would ease pressure on standard handbook provisions — think confidentiality, social media, and civility policies — that many employers rewrote to comply with Stericycle.

"Captive audience" meetings.
In UPS Supply Chain Solutions, the GC has moved to withdraw exceptions in favor of overturning the current Amazon.com Services (2024) rule, which bars employers from requiring employees to attend meetings where the employer expresses its views on unionization. She's urging a return to the decades-old Babcock & Wilcox standard, which permitted mandatory captive-audience meetings. This is a significant one for any employer that uses employee meetings as part of a union-avoidance or communication strategy.

Statements predicting the impact of unionization.
In the same UPS case, GC Carey has explicitly broken from her predecessor's position under Starbucks/Siren Retail (2024) and will instead urge the Board to reinstate Tri-Cast (1985), a more permissive standard for employer statements predicting what might happen to wages, benefits, or working conditions if a union is voted in.

Dress codes.
In Starbucks Corporation, the GC argues against the employee-protective standard from Tesla (2022) and asks the Board to reinstate Wal-Mart Stores (2019), which gave employers more room to enforce dress code and uniform policies — including logo and pin restrictions — without running afoul of Section 7.

Waiver of the right to bargain.
In HPC Industrial Group, the GC has flagged Endurance Environmental Solutions (2024) for reversal and intends to push for a return to the MV Transportation (2019) "contract coverage" standard, which gives more weight to broad management-rights clauses as a basis for unilaterally changing terms and conditions of employment without additional bargaining.

Positions the GC Intends to Raise When the Right Case Comes Along

These are precedents GC Carey has flagged as targets but hasn't yet had a procedural vehicle to formally argue. Employers should watch for these to surface in future litigation:

Bargaining orders without an election.
(Cemex Construction Materials Pacific, 2023): The GC wants the Board to abandon the Cemex framework — which allows a bargaining order to issue without a union election in some circumstances — and return to the pre-Cemex combination of Gissel Packing (1969) and Linden Lumber (1971), which is generally viewed as more protective of an employer's right to insist on a secret-ballot election.

Duty to bargain before changing existing terms.
(Wendt Corporation and Tecnocap, both 2023): These decisions currently require bargaining over changes even where there's longstanding past practice guiding the action. The GC views this as slowing down routine contract administration and wants it revisited.

Union dues checkoff after contract expiration.
(Valley Hospital Medical Center, 2022): The GC wants to return to the 1962 Bethlehem Steel rule, under which an employer's obligation to deduct union dues from paychecks ends automatically when the collective bargaining agreement (and its checkoff clause) expires — rather than continuing post-expiration as Valley Hospital currently requires.

Objector fee disclosures.
(UFCW Local 700/Kroger, 2014): The GC intends to argue that unions should have to disclose more detailed information to dues objectors than Kroger currently requires under Beck and California Saw & Knife Works.

Protected concerted activity and workplace conduct.
(Miller Plastic Products and Lion Elastomers II, both 2023): The GC has specifically called out Lion Elastomers II as extending protection to employee conduct — including conduct that would otherwise be prohibitable — that is only tenuously connected to activity protected under the Act. This case is already pending on remand before the Board.

Enhanced/"make whole" remedies.
(Thryv, 2022): The GC wants the Board to reconsider the expanded consequential-damages remedy adopted in Thryv, noting that courts have repeatedly cut back on it and that it hasn't yet been tested through a full compliance proceeding.

What This Means for Your Organization

Every item above requires the Board to actually rule in the GC's favor. National Labor Relations Act will likely stay in flux for months to come.

Practical Takeaways (For Now)

If your severance agreements, handbook policies, or dress code provisions have been revised in the last two to three years specifically to comply with McLaren Macomb, Stericycle, or Tesla, it may be worth flagging those provisions for a fresh look once the Board rules. This may not mean revising them immediately, but it may help you determine in advance what revisions may be permitted or advisable.

If you would rely on mandatory employee meetings as part of your communications strategy during organizing activity, the captive-audience question is one to watch closely, since a reversal would restore an employer tool that's currently off the table, except in a few states that have outlawed captive-audience meetings under state law.

If you're a unionized employer with a broad management-rights clause, the outcome in the waiver-of-bargaining cases could materially affect how much unilateral flexibility you have when administering the contract.

The core problem is not that AI is involved in employment decisions. It is that AI outputs are increasingly being treated as conclusions rather than inputs, and that shift changes how those decisions look in a deposition, an EEOC position statement, or a jury instruction.

The Employer Cannot Delegate the Decision

Title VII, the ADA, the ADEA, and their state and local counterparts all impose liability on the employer, not on the software the employer purchased. An employer cannot defend a discrimination claim by pointing to a vendor’s algorithm and arguing that a machine, not a person, made the call. Regulators have already made this point explicit. The EEOC has stated that employers remain responsible for adverse impact caused by algorithmic decision-making tools even when a third-party vendor built and maintains the tool. Several state laws, including New York City’s Local Law 144 governing automated employment decision tools, impose independent audit and notice obligations directly on the employer using the tool.

That means the familiar advice to document a legitimate, nondiscriminatory reason for an adverse action now has a second layer. It is not enough that AI flagged the employee. The employer must show that a human reviewed the flag, understood why it was generated, and exercised independent judgment before acting on it. A termination file that says only “system-generated performance score of 2.1, employee terminated” is a much harder file to defend than one that documents what a manager actually observed and considered.

Disparate Impact Hides Inside the Model

Disparate treatment claims require some evidence of intent, but disparate impact claims do not, and AI-driven employment tools are a natural fit for disparate impact theories. A scoring model trained on historical performance or attrition data can quietly reproduce whatever bias existed in that history. A resume screening tool can learn to penalize employment gaps, certain schools, or language patterns that correlate with protected characteristics even though the model was never told to consider race, sex, age, or disability directly.

The practical exposure here is twofold. First, if a plaintiff’s counsel obtains statistical evidence that an AI tool selects or rejects candidates or employees at meaningfully different rates across protected groups, the employer will need validation data showing the tool is job related and consistent with business necessity, the same standard that has applied to any selection device since Griggs v. Duke Power Co. Vendors rarely provide this validation data unprompted, and employers frequently discover during litigation that they never asked for it. Second, an employer that never tested its own tool for adverse impact will have a difficult time arguing it acted reasonably, even where no discriminatory intent existed. Ignorance of how the tool works is not a defense; in a disparate impact case, it can be the plaintiff's best evidence that no one was minding the store.

Discovery Now Reaches Further Than the Personnel File

AI adoption expands what is discoverable in an employment case well beyond the traditional personnel file. Prompts entered by HR or supervisors, model outputs and confidence scores, version histories showing when a scoring model was retrained, and internal communications about why an alert was or was not acted on are all now fair game. Plaintiffs’ counsel is increasingly requesting this material specifically, because it can show not just that an adverse outcome occurred, but what the company knew and when.

This creates a preservation problem many employers have not yet solved. Some platforms overwrite scoring history as new data comes in or retain outputs only briefly by default. If a company has no policy governing retention of AI-generated employment data, it may find itself unable to produce records that plaintiffs assume exist, inviting a spoliation argument, or it may find that the only surviving record is an unfavorable one that a human reviewer never actually relied upon but that now looks, in hindsight, like the smoking gun.

Practical Governance Closes the Gap

None of this counsels against using AI in employment decisions. It counsels against using it without a governance structure built for the way these tools will actually be examined after the fact. At a minimum, employers should be able to show that a person with real authority to disagree reviewed any AI-generated recommendation before it became a personnel action, that the tool has been validated or at least tested for disparate impact on a periodic basis, that managers are trained to document their own independent reasoning rather than simply citing the tool's output, and that retention practices for AI-generated data are deliberate rather than accidental. Where a jurisdiction imposes specific notice, bias audit, or disclosure obligations on automated employment decision tools, those requirements need to be built into the rollout, not addressed after a candidate or employee complains.

The employers best positioned when one of these tools produces a bad outcome are the ones who can show, with contemporaneous documentation, that a human being was accountable for the decision the whole way through. The tool can inform that judgment. It cannot substitute for it, and the law has no intention of letting it try.

Follow the Contract Documents

Most contracts outline a process for declaring breach and terminating the contract. Generally, these steps must be followed. Some contracts identify with specific precision the types of breaches that allow for termination versus other remedies. The governing law (e.g., which state’s law applies) can have significant effect on the righteousness of a termination. Some states rigidly require all processes, notices, and terms set forth in the contract termination clauses to be satisfied prior to termination. Other states may allow exceptions to the drawn-out termination process, affording quick termination that shortcuts contractual notice clauses, depending on the circumstances. Still, whether a contract can be terminated more quickly than set forth in the contract is open to interpretation, and terminating more quickly than the contract strictly requires increases the risk of wrongful termination. Meanwhile, if the contractual processes have been followed, it significantly reduces the risk of a wrongful termination.

Thus, as a general rule of thumb, following the contract processes for declaring breach and termination is a good start.

Best Practice is to Issue a Notice to Cure

Most contracts that outline a process for termination also require issuing a notice to cure prior to termination. As previously stated, it is best to follow the contractual requirements. Even if the contract does not require a notice to cure, it is typically best practice to still issue a notice to cure.

There are two main reasons why this is recommended. First, if the default party cures the breach, then, perhaps there is no need to terminate, because the work has been brought “back on track.”  Generally, forcing the defaulting party to cure the breach at its own cost is less expensive than terminating and fronting the costs to bring in another trade to finish or cure the work.

Second, when proving that the termination was justified, typically it must be demonstrated that the defaulting party was in “material breach” of the contract. If the breach pertains to a critical component of the work, and the defaulting party fails to cure it after a proper notice, that is very strong proof that the defaulting party was in material breach and cannot perform. Thus, the termination for cause is more likely to be adjudged as righteous and justified after a failed cure opportunity. Meanwhile, if no opportunity to cure was provided, the defaulting party can argue that it would have cured the breach. If the breach could have been cured, that is strong evidence that the breach was not material (it was fixable).

Thus, a notice to cure is typically the “other foot dropping” that proves the default could not be cured and therefore the defaulting party was in material breach. Some states require a notice to cure to be issued prior to termination for cause.

Lastly, sometimes the breach has already, previously been cured by the defaulting party. Occasionally, an irritated higher-tier party will provide a lengthy list of transgressions and reasons for termination, but all of them are old, stale, and already cured. Generally, it is problematic to terminate for cause, if the reason for the termination has already been cured. Usually, terminating for cause requires the defaulting party to be in current, uncured breach at the time of termination. If terminating a party for yesteryear’s transgressions, then, you are not technically terminating for breach of contract (it was already cured); instead, you are terminating because you are still angry about it. But that is not a justified basis for termination under the law. It is an uncured (or uncurable) material breach that justifies termination, not a subjective opinion that the party was incompetent due to past issues that are of no current moment.

Document the Breach of Contract, the Remediation, and the Costs/Losses

When terminating a defaulting party, it is important to document breach. Photographs, daily logs/reports, notices to cure, and meeting minutes should corroborate and prove the breach.  Documentation of the redesign or remediation work should be well maintained, including annotated sketches or drawings to explain the details of the breach and remediation. Likewise, clear, segregated cost tracking proving the specific additional costs for remediation and cure of the issue should be maintained.

Also, best practice is to maintain documentation of the curative work to show and explain the steps taken to cure the issue. Often, the curative work itself speaks volumes as to what the problem was.

Preserve Evidence, Allow Access to Evidence, and Avoid Spoliation

Obviously, the evidence of breach, notice, termination, and the remediation work should be preserved. Relatedly, it is best to provide notice of the pending remediation work and allow the terminated party access to the site for a last inspection of the issue prior to the remediation work occurring. This is because, once the remediation work commences, the evidence of breach will inherently be destroyed and manipulated. Sometimes the terminated party will dispute the evidence, and argue that if it had been allowed the opportunity to inspect the defect, it would have been able to prove that the work was in fact satisfactory or a less expensive cure could have been utilized. The best approach is to allow the terminated party to access the site to inspect the work and observe the remediation work. The terminated party cannot interfere with the work or project, of course, but providing reasonable access for inspection (and sometimes destructive testing) is the best practice.

Lastly, sometimes, if allowed to inspect prior to the final termination, the defaulting party might present analysis or evidence to change minds about the course for remediation.

Ensure that All Interested Parties Have Been Given Notice

Sometimes there is a reason to give notice of the termination to third parties. For example, if there is a performance bond posted by the defaulting party, typically it is best to give notice to the surety. Also, sometimes the contract documents require notice of a termination to be given to either a lender, higher tier, or owner.

Consider Whether Statutes Impose a Limitation or Constraint on Termination

Sometimes the basis for termination might conflict with a separate statute. For example, under the Bankruptcy Code it is technically a violation to terminate a contract on the basis of a declared bankruptcy. You must seek bankruptcy court approval for terminating a contract with a bankrupt debtor. Other times, statutory payment acts or other statutes may require a process or steps to be taken prior to termination. This is particularly true if there are withholdings or demands for payment, which is frequently the case.

Ultimately, termination of a party on a construction project is a very strong action with significant repercussions. Missteps in the termination process can compound losses and escalate risk. Care must be taken to approach the termination with careful consideration of strategy and planning in the best interests of both the project and the litigation claims/defenses. It is highly recommended to consult with legal counsel starting with the notice and termination period. Lastly, these approaches are general points for consideration; recognize that each specific situation, project, or contract will have different factors to consider when terminating a downstream party.

JEFFREY C. BRIGHT is a Principal attorney in Offit Kurman’s Construction Practice Group and maintains a multi-state construction law practice, representing contractors, subcontractors, owners, construction managers, design-builders, and design professionals. He is licensed and active in construction law matters in PA, MD, DC, VA, and CA. In addition to handling construction litigation and project disputes, including termination of contracts mid-project, he regularly advises on the preparation, revision, and negotiation of construction contracts for various project delivery systems. He can be reached at jeff.bright@offitkurman.com.