Business Torts: Recent Developments 2025-2026 — Federal vs. State Trade Secrets, AI-Generated IP, and the End of Non-Solicitation

The 2025-2026 period has produced significant shifts in California business tort law across four areas that are reshaping how companies protect proprietary information, compete for talent, and litigate unfair business practices. This annual review examines the developments that California businesses and their counsel need to understand.

I. DTSA vs. CUTSA: Navigating the Federal-State Trade Secret Landscape

Since the Defend Trade Secrets Act of 2016 created a federal civil cause of action for trade secret misappropriation (18 U.S.C. §§ 1831–1839), California litigants have had to navigate the relationship between the DTSA and the California Uniform Trade Secrets Act (Civil Code §§ 3426–3426.11). That relationship has grown more complex, and the strategic calculus for choosing between federal and state claims — or asserting both — has sharpened considerably.

The Non-Preemption Principle

A threshold point that continues to generate confusion: the DTSA expressly does not preempt state trade secret law. 18 U.S.C. § 1838 provides that the Act shall not be "construed to preempt or displace any other remedies, whether civil or criminal, provided by United States Federal, State, commonwealth, possession, or territory law." This means a plaintiff may assert parallel DTSA and CUTSA claims arising from the same misappropriation — and routinely should. The statutes are complementary, not mutually exclusive, and each provides remedies and procedural advantages the other does not.

The Ex Parte Seizure Remedy

The DTSA's most distinctive feature remains its ex parte seizure provision, codified at 18 U.S.C. § 1836(b)(2). In extraordinary circumstances, a court may issue an order for the seizure of property necessary to prevent the propagation or dissemination of a trade secret — without notice to the defendant. CUTSA provides no equivalent remedy. While injunctive relief is available under both statutes, the seizure power is uniquely federal and has been invoked in cases involving imminent departure of key employees with access to highly sensitive information, particularly in the technology and defense sectors.

Courts have applied the seizure provision cautiously. The statute requires the applicant to demonstrate, among other things, that an injunction would be inadequate, that the party against whom seizure is sought actually possesses the trade secret, and that the applicant would suffer immediate and irreparable injury absent seizure. The procedural safeguards — including a mandatory post-seizure hearing within seven days — reflect congressional intent to cabin this extraordinary remedy.

Whistleblower Immunity and Employer Obligations

The DTSA's whistleblower immunity provision (18 U.S.C. § 1833(b)) shields an individual from criminal or civil liability under any federal or state trade secret law for disclosing a trade secret in confidence to a government official or attorney solely for the purpose of reporting or investigating a suspected legal violation. Critically, employers face a practical obligation: any employer that does not provide notice of this immunity in contracts governing trade secrets or confidential information forfeits the right to recover exemplary damages or attorney fees in a DTSA action against an employee. This notice requirement has caught employers off guard, and failure to include the required language in employment agreements, non-disclosure agreements, and separation agreements remains a common oversight.

Forum Selection and Statute of Limitations

Both the DTSA and CUTSA carry three-year statutes of limitations, but they accrue differently. CUTSA's three-year period runs from the date the misappropriation is discovered or should have been discovered (Civil Code § 3426.6). The DTSA uses a similar discovery rule. The practical difference lies in forum: the DTSA provides an independent basis for federal subject matter jurisdiction, allowing trade secret plaintiffs to access federal court without diversity of citizenship. For California companies engaged in disputes with California competitors, the DTSA may be the only pathway to federal court.

The "reasonable measures" requirement is common to both statutes. Under CUTSA § 3426.1(d), a trade secret must derive independent economic value from not being generally known and must be the subject of efforts that are "reasonable under the circumstances" to maintain its secrecy. The DTSA's definition at 18 U.S.C. § 1839(3) tracks this language. In practice, litigating the reasonableness of protective measures — password protections, access controls, confidentiality agreements, compartmentalization of information — remains the single most contested factual issue in trade secret cases under both statutes.

II. AI-Generated Trade Secrets: Defining "Reasonable Measures" in the Age of Machine Learning

The rapid integration of artificial intelligence into business operations has introduced a novel and largely unresolved question in trade secret law: when an AI system generates commercially valuable information by processing proprietary datasets, does the output qualify as a protectable trade secret, and what constitutes "reasonable measures" to maintain its secrecy?

Can AI-Generated Information Be a Trade Secret?

Under CUTSA § 3426.1(d), a trade secret is information that derives independent economic value from not being generally known to or readily ascertainable by other persons who can obtain economic value from its disclosure or use. Nothing in this definition requires that a human being generate the information. An AI system that analyzes proprietary sales data, customer behavior patterns, or manufacturing parameters and produces insights not generally known to competitors may produce output that satisfies the statutory definition — provided the owner takes reasonable steps to maintain secrecy over both the input data and the AI-generated output.

The harder question is whether AI-generated insights that could theoretically be reproduced by anyone with access to sufficient computing power and publicly available training data retain the "not readily ascertainable" quality that CUTSA requires. Where the competitive value derives from the proprietary dataset rather than from the AI model itself, the trade secret protection likely attaches to the dataset and the specific outputs — not to the general methodology.

The Confidentiality Crisis: Third-Party AI Platforms

The most immediate threat to trade secret protection in the AI context is not a legal ambiguity but a practical one: employees inputting confidential business information into third-party AI platforms. The widely reported 2023 Samsung incident — in which employees entered proprietary source code and internal meeting notes into a generative AI chatbot — illustrates the risk. When confidential information is submitted to an external AI service, it may be incorporated into training data, stored on servers outside the company's control, or made accessible to the platform provider's employees. Any of these outcomes can destroy the secrecy element required for trade secret protection.

Courts have not yet squarely addressed whether inputting a trade secret into a third-party AI platform constitutes a failure to take "reasonable measures" to maintain secrecy, thereby forfeiting trade secret status. But the analytical framework is straightforward: if the trade secret owner knows or should know that the platform's terms of service permit retention or use of submitted data for model training, continued use of the platform without adequate safeguards undermines the reasonableness analysis.

Practical Recommendations

Companies seeking to preserve trade secret protection in an AI-enabled environment should implement a multi-layered approach. First, adopt and enforce a written AI acceptable use policy that specifically identifies the categories of information that may not be entered into external AI systems. Second, review and negotiate contractual protections with AI vendors, including data processing agreements that prohibit use of submitted data for model training, require deletion upon request, and impose confidentiality obligations at least as stringent as those in the company's standard NDAs. Third, deploy technical controls — data loss prevention tools, network-level restrictions on AI platform access, and monitoring of outbound data flows — to enforce the policy through means other than employee self-policing. Fourth, conduct regular employee training that goes beyond distributing the policy: employees must understand why entering confidential data into an AI tool can permanently destroy legal protections that took years to build.

Whether AI-generated insights derived from proprietary datasets constitute independently protectable trade secrets remains an open question. The strongest position is to treat the proprietary dataset itself as the trade secret and to classify AI-generated outputs as derivative information subject to the same protective measures. This approach avoids the unsettled legal question of AI authorship and grounds the protection in the established framework of data-as-trade-secret.

III. Non-Solicitation Agreements Voided: The AMN Healthcare Aftermath

California's hostility to post-employment restrictive covenants is not new — Business and Professions Code § 16600 has declared that "every contract by which anyone is restrained from engaging in a lawful profession, trade, or business of any kind is to that extent void" since 1872. But for decades, a significant open question persisted: did § 16600 apply to non-solicitation agreements — provisions restricting a former employee from soliciting the employer's clients or employees — or only to traditional non-compete clauses?

AMN Healthcare and the Elimination of Non-Solicitation

The California Court of Appeal's decision in AMN Healthcare, Inc. v. Aya Healthcare Services, Inc. resolved the question in 2020, holding that § 16600 invalidates non-solicitation provisions that restrain an individual from practicing their profession. The court rejected the argument that customer and employee non-solicitation agreements are categorically different from non-competes, finding instead that any contractual restriction that effectively prevents a person from engaging in their occupation violates § 16600.

This holding built on the California Supreme Court's framework in Edwards v. Arthur Andersen, 44 Cal.4th 937 (2008), which rejected the so-called "narrow restraint" exception that some courts had applied to allow limited non-competes. Edwards established that § 16600 means what it says: every contract in restraint of trade is void, not merely unreasonable restraints. AMN Healthcare extended this categorical approach to non-solicitation clauses.

The 2023 Legislative Reinforcement: AB 1076 and SB 699

The California Legislature codified and expanded the AMN Healthcare holding through two statutes effective January 1, 2024. Assembly Bill 1076 added § 16600.5 to the Business and Professions Code, confirming that § 16600 applies to all restrictive covenants — non-competes, customer non-solicitation, and employee non-solicitation — regardless of whether the agreement was signed in California or another state.

Senate Bill 699 added § 16600.5's enforcement mechanism, making it unlawful for an employer to enter into or attempt to enforce a noncompete agreement that is void under California law, regardless of where and when the agreement was signed. The statute also imposed a notification requirement: employers were required to notify current and former employees, by February 14, 2024, that any non-compete or non-solicitation clause in their agreements was void and unenforceable. Failure to provide this notification exposes employers to claims under § 16600.5 and potential liability under the unfair competition law.

Choice-of-Law Implications

SB 699's extraterritorial reach has created significant choice-of-law friction. A non-solicitation agreement governed by the law of a state that enforces such provisions — such as Florida, Texas, or Massachusetts — is void and unenforceable if the restricted individual works primarily in California or is a California resident. California courts have consistently applied § 16600 notwithstanding contractual choice-of-law clauses selecting another state's law, finding that § 16600 reflects a fundamental public policy of the state.

The practical result: any employer with California-based workers should assume that non-solicitation provisions in their agreements are unenforceable as to those individuals, regardless of where the company is headquartered or what law the agreement selects. The distinction between restricting solicitation and simply hiring former colleagues remains significant — nothing in California law prevents an employer from hiring a competitor's employees, and a departed employee is free to accept inquiries from former clients who seek them out independently.

IV. UCL Section 17200 Enforcement: Intensifying Scrutiny of Unfair Business Practices

California's Unfair Competition Law (Business and Professions Code §§ 17200–17210) has long served as a versatile tool in business tort litigation, and recent developments have reinforced its role as both a complement to and, in some cases, a substitute for traditional tort claims.

The Three Prongs

The UCL prohibits any "unlawful, unfair, or fraudulent business act or practice." Each word defines an independent prong with its own analytical framework. The unlawful prong operates as a "borrowing" statute — a violation of any other law, regulation, or rule can serve as the predicate for a UCL claim. As the California Supreme Court explained in Korea Supply Co. v. Lockheed Martin Corp., 29 Cal.4th 1134 (2003), the UCL's coverage is "sweeping" and allows courts to address practices that violate the policy or spirit of other laws even where those laws do not provide a private cause of action.

The unfair prong has generated two distinct analytical tests depending on the context. For actions between competitors, the California Supreme Court in Cel-Tech Communications v. Los Angeles Cellular Telephone Co., 20 Cal.4th 163 (1999), established that conduct is "unfair" only if it "threatens an incipient violation of an antitrust law, or violates the policy or spirit of one of those laws because its effects are comparable to or the same as a violation of the law, or otherwise significantly threatens or harms competition." In consumer cases, courts have applied a broader balancing test weighing the utility of the defendant's conduct against the gravity of the harm to the plaintiff. This dual-standard framework continues to create analytical challenges, particularly in cases that involve both competitor and consumer dimensions.

The fraudulent prong requires only that the public is likely to be deceived — it does not require proof of actual deception, reliance, or scienter. This lower threshold makes the fraudulent prong a powerful tool in cases where a defendant's representations are misleading but may not meet the heightened specificity requirements of common-law fraud under Civil Code § 1709.

Standing After Proposition 64

Since the passage of Proposition 64 in 2004, private UCL plaintiffs must demonstrate that they "suffered injury in fact and lost money or property as a result of the unfair competition." The California Supreme Court interpreted this requirement in Kwikset Corp. v. Superior Court, 51 Cal.4th 310 (2011), holding that a plaintiff must show a causal connection between the alleged UCL violation and the economic injury, but that this standard is not as demanding as common-law standing. In business tort contexts, the standing requirement is typically met where the plaintiff lost revenue, customers, or market share as a result of the defendant's unfair practices.

Remedies: Restitution and Injunction, Not Damages

A critical limitation of the UCL is its remedial scope. Under § 17203, courts may grant restitution and injunctive relief, but the UCL does not authorize damages, punitive or otherwise. Korea Supply clarified that "restitution" under the UCL means the return of money or property that the defendant took from the plaintiff through unfair means — it does not encompass disgorgement of profits that the defendant earned from third parties. This limitation often determines whether a UCL claim serves as the primary vehicle for relief or merely as a supplement to contract and tort theories that do authorize full compensatory and punitive damages.

Despite this limitation, UCL claims have become increasingly valuable in business tort litigation. They allow plaintiffs to challenge conduct that may not fit neatly into traditional tort categories, provide injunctive relief that can halt ongoing unfair practices, and — through the borrowing function of the unlawful prong — create private enforcement mechanisms for statutes that do not include them.

The 2025-2026 developments surveyed here reflect a California business tort landscape that is becoming simultaneously more protective of employee mobility and more demanding of companies that seek to protect proprietary information. Navigating these dual pressures requires integrated legal strategies that account for overlapping federal and state frameworks, emerging technology risks, and the expanding reach of unfair competition law.

This analysis is for informational purposes only and does not constitute legal advice. The legal landscape is evolving, and readers should consult qualified counsel for advice specific to their circumstances. No attorney-client relationship is formed by reading this publication.