The Rutgers Law Record | The First Digital Law Journal

BEYOND SELF-REPORTING: VERIFYING PLATFORM INTRANSPARENCY UNDER THE EU DIGITAL SERVICES ACT

Posted on August 4, 2026 by Alessia Zornetta

The European Union’s Digital Services Act (DSA) represents an ambitious attempt to improve the oversight of digital platforms through a multi-layered transparency framework. At its core, the regulation aims to increase platform accountability through strict transparency obligations and independent verification mechanisms. This article examines a key element of this framework: the role of independent auditors in verifying platform transparency and how it functions alongside other DSA transparency tools to address the reliability problem. While the DSA introduces various tools to improve platform oversight, the auditing requirements under Article 37 represent one of its most significant innovations in addressing the longstanding challenge of unverifiable platform disclosures.

For transparency to be truly meaningful in platform governance, it must be both reliable and independently verifiable. Platform transparency reports have long faced criticism for serving more as public relations tools than genuine accountability mechanisms, with their accuracy and completeness frequently called into question. This reliability problem has led to sustained calls for the development of verification mechanisms for platform data.

This article provides, to the best of the author’s knowledge, the first academic analysis of the recently published DSA audit reports, contributing insights on how the auditing mechanism is functioning in practice. Through an examination of the audit reports from nineteen Very Large Online Platforms and Search Engines, the article reveals significant challenges in the verification of transparency practices, identifying patterns across different platforms that have not previously been documented in academic literature. While some initial blog posts and analyses have highlighted selected aspects of these reports, this article offers a systematic examination of how auditors approached transparency verification, what compliance issues they identified, and what these findings tell us about the reliability of platform transparency more broadly.

The underlying assumption driving demands for verification is that transparency reports could become effective tools for public policy decisions about platforms’ abilities to limit online harm, but only if there exists a reliable way to verify the platforms’ representations of their content moderation efforts accurately reflect reality. The role of auditors in the financial sector is often pointed at as an example of what should happen with online platforms. Just as financial auditors examine banks’ procedures for monitoring suspicious activities rather than reviewing each transaction, platform auditors could verify content moderation transparency by evaluating the processes that generate reported outcomes. However, unlike financial auditing, which operates within well-established standards and methodologies developed over decades, platform auditing has emerged in a regulatory vacuum where platforms use different reporting formats and metrics, making meaningful cross-platform comparison and verification quite challenging.

View Entire Article:

https://lawrecord.com/wp-content/uploads/2026/08/53_Rutgers_L_Rec_270-Zornetta.docx-2.pdf

Share and Enjoy:

BETTER THAN REVENGE: A COMPARATIVE ANALYSIS OF NEW YORK AND NEW JERSEY’S LEGAL REMEDIES FOR REVENGE PORN AND AI-GENERATED EXPLICIT CONTENT

Posted on May 7, 2026 by Courtney Stonefelt

53 Rutgers L. Rec. 240 (2026) | WestLaw | LexisNexis | PDF

On March 15, 2022, the mother of an anonymous woman was contacted through social media private message.^[1] When opening the message, the mother was confronted with an image of her naked daughter, “breasts and genitals in full view.”^[2] The images were sent from the anonymous woman’s (the plaintiff) ex-lover.^[3] The messages included “outside links containing additional images and videos of [the] plaintiff” which were sent to both the “plaintiff’s mother [and] business associates” of the woman.^[4] This anonymous woman’s story represents one instance of an increasingly prevalent form of sexual abuse that has devasted countless lives across the United States. Recent statistics from the Cyber Civil Rights Initiative reveal that one in twelve adult Americans have been victims of nonconsensual distribution of intimate images and that one in eight adult social media users have been targets of nonconsensual distribution of intimate images, or nonconsensual pornography.^[5]

* J.D. Candidate, Rutgers Law School, 2026.

^[1] P.F. v Brown, 2024 NY Slip Op 51356 (U), *1–2 (N.Y. Sup. Ct., Queens County 2025).

^[2] Id.

^[3] Id.

^[4] Id. (internal quotations omitted).

^[5] See generally Cyber Civ. Rts. Initiative, https://cybercivilrights.org/ (last visited Mar. 17, 2025).

View The Entire Article

Share and Enjoy:

PROTECTION TO “INFINITY AND BEYOND”: WHY FORMER EMPLOYEES ARE NOT PROTECTED BY THE ANTI-RETALIATION SAFEGUARD OF THE FALSE CLAIMS ACT

Posted on April 3, 2026 by Michael Giorgio

53 Rutgers L. Rec. 219 (2026) | WestLaw | LexisNexis | PDF

0. ABSTRACT

The False Claims Act (FCA) is a significant piece of federal legislation enacted to prevent individuals or companies from engaging in fraudulent activities to fleece government programs. Whistleblowers play a critical role in reporting these fraudulent actions. Unfortunately, whistleblowers often become targets of retaliatory actions either during or after their employment. The FCA’s 1986 amendments provided protections to whistleblowers, who are identified as “employees.” The Sixth and Tenth Circuits have disagreed on who is considered an “employee.” The Sixth Circuit has agreed to extend whistleblower protection to former employees whereas the Tenth Circuit has refused to do so.

Part I of this note introduces the FCA, including the unique qui tam provision and relevant statistics of the Act. Part II explains the history and background of the Act and provides detailed reasonings why the circuit courts disagree as to the scope of the term “employee.” Part III provides justifications why the Sixth Circuit’s extension of protection to former employees is problematic. This includes expounding upon how the Sixth Circuit mistakenly determined the anti-retaliation provision was ambiguous, overly relied on overtly broader precedent and ignored legislative intent,

and failed to use statutory canons to guide its analysis. Part IV addresses counter arguments to extending protection, including the reasons that it would reduce reporting fraud and how most whistleblower protection provisions in other federal statutes are frequently broadly examined by courts. Finally, Part V attempts to provide a new approach to interpreting the word “employee,” including using the economic reality test, which would verify an “employee” as a person who is economically beholden to another.

View The Entire Article

Share and Enjoy:

Ready Player Two … for Tort Liability in the Metaverse?

Posted on February 17, 2026 by Nanci K. Carr et al.

53 Rutgers L. Rec. 193 (2026) | WestLaw | LexisNexis | PDF

In the past few years, augmented reality (AR) and virtual reality (VR) platforms have seen
rapid expansion. The metaverse uses AR and VR to create a parallel virtual world where
users can work, play, and interact. With one of the world’s largest tech giants pivoting to
the metaverse–described as an $800 billion market opportunity–new advances are bound
to make the metaverse more interactive and immersive than it already is. Developments in
the field of AR and VR that might add sensory cues to create highly immersive and realistic
environments have the potential to evoke real-world stimuli and a complete sense of
presence and bodily embodiment in the virtual environment. In a highly immersive and
realistic virtual environment, if a harmful event occurs, users may receive real-world
sensory cues, including taste, smell, and touch, and react to those events, actively
experiencing and embodying the harm. These environments pose compelling legal and
ethical questions for practitioners and law students alike. This article proposes integrating
AR- and VR-based case studies into legal education to enhance student engagement and
critical thinking, illustrated by the hypothetical case of Olivia v. Thomas442, in which a
VR user experiences emotional and physical distress following a virtual attack. The article
demonstrates the pedagogical value of such hypothetical scenarios in preparing future
practitioners for emerging challenges in technology-driven contexts.

View The Entire Article

Share and Enjoy:

#DISCLOSED? NAVIGATING THE REGULATORY DIVIDE IN SOCIAL MEDIA MARKETING

Posted on December 23, 2025 by Madison Staffordsmith

53 Rutgers L. Rec. 159 (2025) | WestLaw | LexisNexis | PDF

Checking your phone has likely become a seamless part of your daily routine, often within the first ten minutes of waking up.^[1] While you catch up on your various social media applications, you are exposing yourself to numerous advertisements, possibly without even knowing.^[2] Perhaps the advertisement is shared from an account you trust based off the mere fact you feel connected to the person posting it.^[3] That feeling of trust and connection combined may lead you to become a consumer of the product or item that is being promoted, without taking much convincing.^[4]

Consumers of media have been subject to multiple forms of influencing long before the term ‘influencer’ “became an entry in modern lexicons.”^[5] Throughout history, influential figures have utilized traditional media forms such as newspapers, radio, and television to sway public opinion.^[6]

Today, brands increasingly capitalize on the prominence and influence of social media influencers to drive consumer purchases and boost awareness of their brand.^[7] Posting on social media may once have been seen merely as a recreational activity for all, but it has now transitioned into a professional career choice for some.^[8] The continuous rise of social media combined with the younger generational views toward the workforce shifts social media influencing into what can be viewed to be the new American Dream.^[9]

The Federal Trade Commission (“FTC”) considers “a social media influencer as an individual with a large social media following” whom brands contact to promote products or services.^[10] With the rapidly evolving and growing industry, several international bodies have begun practices for regulation.^[11] Despite having regulatory agencies around the world, such as the FTC and the Australian Competition and Consumer Commission (“ACCC”), influencers “continuously fail to disclose advertisements according to their country’s laws and guidelines.”^[12] As a result, influencers, and brands they work with, may engage in misleading or unfair practices that hinder consumers from making well-informed choices.^[13]

This Note explores the United States and Australia’s contrasting regulatory approaches to social media influencer marketing. Part II discusses the development and evolution of consumer protection laws in both the United States and Australia, highlighting how these distinct legal histories have shaped their current regulatory frameworks. Part III examines the current enforcement strategies, contrasting the FTC’s reactive, case-by-case approach with the ACCC’s proactive monitoring system, and analyzes their effectiveness in the international digital marketplace. Part IV explores the challenges to regulatory convergence, including constitutional frameworks, market scale disparities, and technological innovations that complicate oversight efforts. Finally, Part V considers future trajectories in influencer regulation as emerging technologies, platform evolution, and changing consumer behaviors continue to transform the digital marketplace.

View The Entire Article

* J.D. Candidate, Rutgers Law School, 2026.

^[1] John Shumway, Study finds nearly 57% of Americans admit to being addicted to their phones, CBS News Pittsburgh (Aug. 30, 2023), https://www.cbsnews.com/pittsburgh/news/study-finds-nearly-57-of-americans-admit-to-being-addicted-to-their-phones/ (reporting that 89% of people check their phones within the first ten minutes of waking up).

^[2] Thinking vs. Feeling: The Psychology of Advertising, USC Dornsife (Nov. 17, 2023), https://appliedpsychologydegree.usc.edu/blog/thinking-vs-feeling-the-psychology-of-advertising/ (explaining how brands are marketing themselves with advertisements on social media).

^[3] Consumers Seek Influencers Who Keep It Real, Matter (Feb. 22, 2023), https://www.matternow.com/blog/consumers-seek-influencers-who-keep-it-real/ (finding that 69% of respondents are likely to trust a friend, family member or influencer recommendation over information coming directly from a brand).

^[4] Id.

^[5] When Did Influencers Become a Thing? A Timeline, Influence Insider, https://influence-insider.com/when-did-influencers-become-a-thing-a-timeline/ (noting that the transformative period in the history of influences can be pinpointed to the emergence of platforms like MySpace and the proliferation of the blogosphere in the early 2000s).

^[6] See id.

^[7] Kristin Hovie, That’s Hot: Influencer Stars Should Not Be Blind to Properly Disclosing Social Media Advertisements, 46 Suffolk Transnat’l L. Rev. 71 (2023) (citing Megan K. Bannigan & Beth Shane, Towards Truth in Influencing: Risks and Rewards of Disclosing Influencer Marketing in the Fashion Industry, 64 N.Y.L. Sch. L. Rev. 247, 249 (2019) (describing development of sponsored content on social media sites)).

^[8] Katherine Hu, The Influencer Economy Is Warping the American Dream, The Atlantic Daily (Apr. 18, 2023), https://www.theatlantic.com/newsletters/archive/2023/04/social-media-influencers-american-economy/673762/.

^[9] Id.

^[10] Craig Ganter, Disclosing Under The Influencer: How the FTC Fails to Guide Advertisers and Protect Consumers In Social Media “Influencer” Marketing, Rutgers U.L. Rev. Comments. 47 (2019); see also Lauryn Harris, Comment, Too Little, Too Late: FTC Guidelines on “Deceptive and Misleading” Endorsements by Social Media Influencers, 62 How. L.J. 947, 955-57 (2019) (describing categories of social media influencers recognized by scholars).

^[11] Cheat Sheet For Influencer Regulation in 16 Different Markets, Hello Partner (May 18, 2020), https://hellopartner.com/2020/05/18/influencer-regulation-different-markets/.

^[12] Hovie, supra note 7, at 74.

^[13] Id.

Share and Enjoy:

THE EVIDENTIARY ADMISSIBILITY OF EXTREME WEATHER EVENT ATTRIBUTION OPINIONS

Posted on November 21, 2025 by Matthew Wickersham and Briana Matusovsky

53 Rutgers L. Rec. 125 (2025) | WestLaw | LexisNexis | PDF

A new field of climate science seeks to link specific weather events—like hurricanes, wildfires, or heatwaves—to human-caused climate change. These types of expert opinions have often been touted as a means to support climate change litigation that seeks compensation for the emissions of greenhouse gases from specific companies or industries, or that challenges inaction by governmental entities in regulating greenhouse gas emissions. These “attribution opinions” are intended to show that a specific plaintiff has been harmed because of a specific extreme weather event and so has standing to seek relief or compensation for this harm. As most climate change cases have been resolved to date on justiciability concerns or other procedural grounds, there has been little analysis as to whether these types of attribution opinions would survive an evidentiary challenge as to their admissibility.

In addition to the standard set by the Federal Rules of Evidence, state courts have adopted a variety of different tests for determining the admissibility of expert opinions. However, there are common principles that are generally applicable. By examining the methodologies identified in published articles from prominent purveyors of attribution science, this article examines whether these types of extreme weather event attribution opinions would be considered to meet these admissibility standards for expert opinions. The article concludes that, as currently described in the scientific literature, these attribution opinions would have difficulty satisfying the evidentiary standards for admissibility.

View the Entire Article

Share and Enjoy:

SOCIAL MEDIA ACCOUNTS AS PROPERTY: A HISTORICAL ANALYSIS OF PERSONAL PROPERTY CLAIMS OVER PLATFORM ACCOUNTS

Posted on October 28, 2025 by Sydney Rose

53 Rutgers L. Rec. 94 (2025) | WestLaw | LexisNexis | PDF

La Baguette, a retail bakery, relied on its Facebook page to reach customers.[1] La Baguette’s Facebook page had 4,000 followers to whom the business communicated special promotions and from which it took customer orders.[2] The business hired two employees to run its social media accounts.[3] Those employees later changed the name of La Baguette’s Facebook page to advertise their own competing bakery, Tito & Tita Langley.[4] The employees refused to provide La Baguette with its own login credentials and therefore forced La Baguette to create a new Facebook page, this time with less than three hundred followers.[5] Tito & Tita Langley’s hijack of La Baguette’s Facebook page allowed them to divert a significant amount of customer orders from La Baguette.[6]

Unfortunately, La Baguette’s story is not unique.[7] Digital platform accounts drive significant value for businesses of all sizes. As of late 2023, 95% of small businesses in the United States used at least one digital platform.[8] Public companies that use digital platforms create “much more shareholder value” than businesses who have minimal or no digital presence.[9] The value to and dependence of businesses on digital platforms is unique in that it is not derived by an asset owned by the business. This can leave businesses and individuals vulnerable.

This article explores, through lawsuit tracking, how courts’ reasoning regarding property rights in digital assets on platforms, like social media accounts, has transformed from the 1990s to current day. This is ultimately to consider the question: what does it mean to have a property right in an online account? What interest do users have in their accounts if the platform ceases operation?

This issue brings to light the implications of intermediary failure. The rise of online intermediaries has created a layered market structure where the rights and existence of all platform user accounts depend on the platform itself. Another example of this phenomenon is non-fungible tokens (“NFTs”). Congress is considering a bill which would prevent NFTs from being considered a security.[10] This would create a personal property interest in the NFT’s owner by putting NFTs in the same class as art, music, literary works, intellectual property, collectibles, and merchandise.[11] But what would happen if the infrastructure that supports the NFT, the blockchain, fails? What personal property interest would the NFT owner have left? The answer may be only a string of code.

While users may feel that their accounts and content expressed on the Internet are their own, the legitimacy of any claim to ownership is contested and contingent on the parties to the particular ownership interest inquiry. Platforms have a superior right to ownership of accounts as against users. However, ownership rights between users are less established.

This paper first documents the rise of online platforms and examines how this evolution impacted courts’ recognition of users’ property rights in their online accounts. Next, this paper traces twelve specific, illustrative rulings in three phases of the Internet: Early Internet: 1990-2005, The Rise of Platforms: 2005-2015, and Contemporary Internet: 2015-Present. This paper then synthesizes a set of default rules which courts have created through the caselaw as they apply common law property principles to digital accounts. Finally, this paper puts these pieces together to describe the layered ownership structure created by digital platforms.

*Sydney Rose earned her J.D. from Southern Methodist University Dedman School of Law in the Spring of 2025, and is currently a practicing attorney at Winston & Strawn.

[1] Pan 4 Am., LLC v. Tito & Tita Food Truck, LLC, No. DLB-21-401 (D. Md. Mar. 3, 2022) (unpublished mem. op.) (noting La Baguette’s reliance on its Facebook page).

[2] Id.

[3] Id.

[4] Id. (noting the Facebook name change to “Tito & Tita Langley”).

[5] Id.

[6] Id.

[7] See infra Bearoff v. Craton, 350 Ga. App. 826, 840-41 (Ga. Ct. App. 2019); see also Int’l Bhd. of Teamsters Loc. 651 v. Philbeck, 464 F. Supp. 3d 863, 872 (E.D. Ky. 2020); see also JLM Couture, Inc. v. Gutman, No. 1:20-cv-10575, ECF No. 431 at 26 (S.D.N.Y. 2023) (amended opinion) (unpublished opinion); see also In re Vital Pharmaceutical 652 B.R. 392, 405 (S.D. Fl. 2023).

[8] Empowering Small Business: The Impact of Technology on U.S. Small Business, U.S. Chamber of Com. (Sept. 14, 2023), https://www.uschamber.com/assets/documents/The-Impact-of-Technology-on-Small-Business-Report-2023-Edition.pdf.

[9] Eric Lamarre et al., The Value of Digital Transformation, Harv. Bus. Rev. (July 31, 2023), https://hbr.org/2023/07/the-value-of-digital-transformation.

[10] Mauro Wolfe & Vincent Nolan, NFT Bill Needs Refining to Effectively Regulate Digital Assets, Duane Morris: Bylined Articles (Feb. 27, 2025), https://www.duanemorris.com/articles/nft_bill_needs_refining_effectively_regulate_digital_assets_0225.html.

[11] H.R. 10544, 118th Cong. (2024) (formerly called the New Frontiers in Technology Act).

View The Entire Article

Share and Enjoy:

ARTIFICIAL AUTHORITY: FEDERALISM, PREEMPTION, AND THE CONSTITUTIONAL STRUCTURE OF AI REGULATION

Posted on October 17, 2025 by Donovan Estrada

53 Rutgers L. Rec. 36 (2025) | WestLaw | LexisNexis | PDF

Executive Order 14,179 (“EO 14,179”), signed by President Donald J. Trump in January 2025, heralds a new federal approach to artificial intelligence (“AI”) governance focused on deregulation and national competitiveness. This Article analyzes EO 14,179’s sweeping changes – notably its revocation of President Biden’s AI executive order (“EO 14,110”) and its directive to produce America’s AI Action Plan–and contrasts them with emerging state-level AI regulations. The July 2025 AI Action Plan emphasizes deregulation, infrastructure expansion, and international competition, even directing federal agencies to consider withholding funds from states enacting burdensome or restrictive AI laws.^[1] Such measures set the stage for a federalism clash with states like Colorado, which passed a landmark Colorado AI Act (SB 24-205) to regulate “high-risk” AI systems, which becomes effective February 1, 2026.^[2] This Article explores the resulting legal tensions under the Spending Clause, Dormant Commerce Clause, and preemption doctrines. It argues that the Administration’s attempt to preempt or penalize state AI regulations by executive fiat raises constitutional red flags under the Spending Clause and tests the limits of executive authority. Simultaneously, state laws like Colorado’s invite scrutiny under Dormant Commerce Clause jurisprudence as potential burdens on interstate commerce. The analysis reviews these constitutional dimensions, including the applicability of Spending Clause constraints and Dormant Commerce Clause precedents, and examines whether federal preemption could override state AI laws. Finally, the Article offers a balanced policy discussion weighing the imperative of innovation and AI leadership against the need for risk mitigation and accountability.

I. Introduction

In early 2025, the United States government dramatically pivoted its approach to AI governance. Upon taking office, President Donald Trump issued Executive Order 14,179 titled “Removing Barriers to American Leadership in Artificial Intelligence,” (“EO 14,179”) signaling a decisive shift toward deregulation and rapid innovation.^[3] EO 14,179 explicitly revoked prior federal AI policies deemed impediments to innovation – most notably rescinding President Biden’s October 2023 executive order on the “Safe, Secure, and Trustworthy Development and Use of AI” (“EO 14,110”).^[4] In its place, EO 14,179 set a national policy of sustaining American “global AI dominance” and directed the creation of a comprehensive federal AI Action Plan to accelerate U.S. AI leadership.^[5]

This federal push for unfettered AI development soon met resistance at the state level. As Washington promoted a light-touch regulatory stance, several states had begun crafting their own rules to address emerging risks perceived in AI. For example, in May 2024, Colorado became one of the first states to enact a broad AI governance law, Senate Bill 24-205, known as the Colorado Artificial Intelligence Act (“Colorado AI Act” or “CAIA”).^[6] Set to take effect on February 1, 2026, the Colorado AI Act imposes transparency, fairness, and accountability obligations on “high-risk” AI systems used in “consequential decisions” like employment, lending, or healthcare.^[7] Colorado’s law–and similar initiatives in states such as Utah and draft proposals in California – reflect growing concern over “algorithmic discrimination” and other AI caused harms in the absence of federal regulation.^[8]

This divergence between a deregulatory federal agenda and proactive state regulations has teed up a classic federalism fight, this time over AI. To be sure, the Trump Administration’s America’s AI Action Plan, released in July 2025 pursuant to EO 14,179, not only lays out a national strategy favoring innovation and infrastructure, but also pointedly targets state laws that try to regulate AI as potential “barriers” to progress.^[9] Trump’s AI Action Plan recommends that federal agencies consider a state’s AI regulatory climate when allocating discretionary funds, and to limit funding if state regulations are deemed “unduly restrictive.”^[10] It also directs the Federal Communications Commission (“FCC”) to evaluate whether state AI rules interfere with federal mandates, hinting at possible preemption efforts.^[11] These measures invert the usual federalism model – instead of enticing states to raise standards through funding, the federal government is pressuring states not to regulate AI in hopes that deregulation will spur innovation.^[12]

The collision course is set: a deregulation-first federal policy versus state-level proactive risk regulation. This Article examines the constitutional and legal implications of this conflict. Part I provides background on Executive Order 14,179 and its corresponding AI Action Plan. Part II discusses Colorado’s AI Act as a case study in state AI regulation and its potential burden on interstate commerce. Part III analyzes the conflict through constitutional lenses – the Spending Clause’s limits on conditioning federal funds, the Dormant Commerce Clause’s constraints on state laws affecting interstate commerce, and principles of federal preemption and executive power. Part IV offers a policy analysis, weighing the benefits of innovation and national uniformity against the values of experimentation and public protection. The Article concludes by considering paths forward to reconcile innovation with governance, positing that a balanced national framework may be needed to avoid protracted federal-state conflict in the AI arena.

^[1]See Winning the Race: America’s AI Action Plan, White House Office of Sci. & Tech. Pol’y (July 23, 2025), https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf (hereinafter ‘America’s AI Action Plan’).

^[2]An Act Concerning Consumer Protections in Interactions with Artificial Intelligence Systems, ch. 198, §§ 1–10, 2024 Colo. Sess. Laws 198 (enacted May 17, 2024, to be codified at Colo . Rev . Stat . § 6-1-1701 et seq. (eff. Feb. 1, 2026)).

^[3] See Exec. Order No. 14,179, 90 Fed. Reg. 8741 (Jan. 23, 2025).

^[4] Exec. Order No. 14,110, 88 Fed. Reg. 75191 (Oct. 30, 2023) (revoked by Exec. Order No. 14,148, 90 Fed. Reg. 75192 (Jan. 20, 2025)).

^[5] Exec. Order 14,179, supra note 3, at § 2.

^[6] An Act Concerning Consumer Protections in Interactions with Artificial Intelligence Systems, ch. 198, §§ 1–10, 2024 Colo. Sess. Laws 198 (enacted May 17, 2024, to be codified at Colo. Rev. Stat. § 6-1-1701 et seq.) (eff. Feb. 1, 2026)).

^[7] Id.

^[8] Artificial Intelligence Policy Act, S.B. 149, 2024 Gen. Sess. (Utah 2024) (enacted March 13, 2024) (codified at Utah Code Ann. § 13-2-12)(establishing transparency and disclosure requirements for AI interactions); Cal. Assemb. B. 331, 2023–24 Leg., Reg. Sess. (Cal. 2024) (proposing oversight for automated decision systems in sensitive contexts).

^[9] See America’s AI Action Plan , supra note 1, at 1.

^[10] See id. at 3.

^[11] Id.

^[12] See South Dakota v. Dole, 483 U.S. 203, 211 (1987) (upholding a conditional highway‑funding incentive for states to adopt a minimum drinking age of 21).

View the Entire Article

Share and Enjoy:

NEW JERSEY PRODUCTS LIABILITY: ANALYZING AMBIGUITIES WHICH ARISE DURING PRODUCTS LIABILITY ACTIONS REGARDING COUNTERFEIT DRUGS

Posted on October 9, 2025 by Dylan Kuncken

53 Rutgers L. Rec. 1 (2025) | WestLaw | LexisNexis | PDF

On December 6, 2023, the family of a 32-year-old man from Toms River, New Jersey found him unexpectedly deceased.^[1] The man’s death was traced back to his usage of a supplement referred to as “gas station heroin,” more formally marketed as “Neptune’s Fix Elixir,” which he purchased from a market store in Point Pleasant Borough, New Jersey.^[2] The primary ingredient within the “Elixir” is tianeptine, a drug that is unregulated by the U.S. Food and Drug Administration (FDA).^[3]

As a result of consuming this product, the man suffered a “seizure, heart failure, cerebral anoxia, and other medical issues” leading to his death.^[4] In outwardly recalling the product^[5], the FDA has emphasized that tianeptine can interact with other medicines in a “life-threatening way,” yet the marketing of the product is not displayed as such.^[6]

This exemplifies only one situation where a product containing tianeptine has caused a consumer to experience adverse reactions, but there are several others that have occurred since 2023.^[7] With many consumers who are harmed from scenarios like the abovementioned seeking relief via a product liability action, they will need to know which parties can be liable for their conduct. New Jersey product liability law has ambiguities of when sellers and/or distributors [within this context] can be held liable for their products and/or conduct. This article will further highlight the epidemic of counterfeit drugs harming consumers across New Jersey and will analyze the ambiguities of the state’s product liability law for this niche issue.

Counterfeit drugs have found their way into the hands of several innocent consumers across the nation.^[8] This has occurred throughout both the pharmaceutical industry and amongst drugs sold in convenience stores.^[9] Many individuals who have suffered cognizable harm from consuming a product that did not contain what they thought/were told it did seek civil damages and are unsure where to start.^[10]

In New Jersey, victims sustaining harm due to consumption or usage of a defective product are permitted to bring an action against the manufacturers, sellers or distributors of that product.^[11] Whether a defendant is liable in products liability is circumstantial, as many aspects of law can be.^[12] Those circumstances can include the manufacturer or sellers’ knowledge of the product being defective, their contribution to the defect(s), the nature of the product, the nature of the defect(s), labeling/failure to warn, etc.^[13]

Under New Jersey Products Liability Law, manufacturers and sellers have a duty to patrons to make or sell a product that is “reasonably safe.”^[14] The term “reasonably safe” means that the product is “reasonably fit, suitable and safe for intended and foreseeable usage.”^[15] In addition, manufacturers and sellers owe this duty only to foreseeable consumers and users of the product.^[16]

It is significantly more difficult for a manufacturer to escape a products liability action than it is for a seller, as sellers at minimum have the defense of providing that they were an “innocent seller.”^[17] The innocent seller defense immunizes a seller of a defective product if they can sufficiently prove that they were unaware of the defect and that they “should not have been aware” of said defect.^[18] If a seller’s innocence defense is insufficient, they can be subject to strict liability.^[19] Within New Jersey law, there are complexities as to when sellers, specifically, of counterfeit drugs, are strictly liable or immunized from liability, as the nature of the defects differentiate.^[20]

This article will address the following: (1) The difference between manufacturers, sellers and distributors under New Jersey Products Liability law; (2) the innocent seller defense and its applicability to the sale of counterfeit drugs under New Jersey law; (3) the applicability of strict liability under New Jersey law to sellers where there is a sale of counterfeit drugs; (4) identification of situations where this issue is applicable; and (5) ananalysis of the complexities of which law is applicable and provide solutions to create a more uniform enforcement of the this issue within New Jersey Law.

[1]*J.D. Candidate, Rutgers Law School, 2026.

See Anthony G. Attrino, N.J. Man, 32, died after consuming Neptune’s Fix Elixir from local market, lawsuit says, N.J.Com (Sep. 13, 2024 at 9:08 A.M.), https://www.nj.com/ocean/2024/09/nj-man-32-died-after-consuming-neptunes-fix-elixir-from-local-market-lawsui

[2] I d .

[3] Id .

[4] Id .

[5] FDA warns consumers not to purchase or use any tianeptine prod. due to serious risks, fda (Oct. 1, 2024 ), https://www.fda.gov/drugs/drug-safety-and-availability/fda-warns-consumers-not-purchase-or-use-any-tianeptine-product-due-serious-risks.

[6] See Attrino, supra note 1.

[7] See generally Fred Charatan, Fake Prescription Drugs are Flooding the U.S., 322 Brit. Med. J. 1446 (Jun. 16, 2001), https://pmc.ncbi.nlm.nih.gov/articles/PMC1173338/. While this article is from 2001, the flooding of prescription drugs referenced within illustrates how many individuals are experiencing adverse effects of counterfeit drugs across the country, and that number has only increased since 2001. Several individuals, especially in the Northeast, are suffering adverse reactions to misrepresented defective/dangerous products such as Neptune’s Fix Elixir.

[8] Id .

[9] Id. (emphasizing the increase in counterfeit pharmaceutical drugs); see also AnneMoss Rogers, Gas Station Heroin, and Other Not-Yet-Regulated Drugs, Mental Health Awareness Educ. (Jan. 26, 2024), https://mentalhealthawarenesseducation.com/gas-station-heroin-and-other-not-yet-regulated-drugs/ (highlighting the rise of unregulated/counterfeit substances being sold at convenience stores and gas station markets). Specifically, with the convenience store substances, they are highly addictive and do not provide adequate consumers. Often, typical consumers of these products are teens and young adults. Counterfeit drugs at convenience stores are highly predatory. Unregulated substances like that highlighted in the referenced article (which will be analyzed later on) are not yet regulated and are often marketed as “mood enhancing,” “dietary supplements” or “focus aids.” Both types of counterfeit substances referenced within are highly dangerous and the question presented in this piece will apply to both.

[10] See generally Who Can be Responsible for the Sale of Counterfeit Medications?, Shapiro Legal Group, PLLC (last visited Sep. 19, 2025), https://www.shapirolegalgroup.com/who-can-be-responsible-for-the-sale-of-counterfeit-medications.html. The referenced legal article underscores the importance of victims of this issue receiving justice and outlines the steps of who can be liable in a legal action. Specifically, it states “counterfeit medications have become increasingly prevalent, putting countless patients at risk. These fake drugs may leave consumers vulnerable to serious health consequences. In these situations, justice must be served…They [counterfeit drugs] may contain toxic or harmful ingredients that can cause adverse reactions, allergic responses, or even poisoning.”

[11] See N.J. Rev. Stat. § 2A:58C-2 (2024).

[12] See N.J. Rev. Stat. § 2A:58C-3 (2024).

[13] See id .

[14] See Model Civ. Jury Charge § 5.40A, Products Liability – Introduction: Caveats to Judges at 2.

[15] See id.

[16] See id .

[17] See N.J. Rev. Stat. § 2A:58C-9(b) (2024).

[18] See Fabian v. Minster MacH. Co. 258 N.J. Super. 261, 272 (N.J. Super. Ct. App. Div. 1992).

[19] See N.J. Rev. Stat. § 2A:58C-9(c) (2024).

[20] See Prod. Liab. Claims: How N.J. L. Protect Consumers, DRAZIN AND WARSAW P.C. (last visited Oct. 6, 2024), https://www.drazinandwarshaw.com/blog/product-liability-claims-how-new-jersey-laws-protect-consumers/.

View the Entire Article

Share and Enjoy:

CAN FEDERATED LEARNING SOLVE AI’S DATA PRIVACY PROBLEM?: A LEGAL ANALYSIS

Posted on May 29, 2025 by Warren B. Chick and Florian Gamper

52 Rutgers L. Rec. 252 (2025) | WestLaw | LexisNexis | PDF

The speed of development in Artificial Intelligence (AI) in recent years has been breathtaking. Yet this comes with its own set of problems.[1] One of these problems is that the current iteration of AI requires vast quantities of data for the training of AI systems, and, currently, the demand for data is outstripping the supply of data.[2] This may hinder the further development and improvement of AI systems. This is sometimes referred to as AI’s data problem.[3] However, making more data available to train AI raises various concerns, not least, in relation to data privacy, thus, AI has a data privacy problem.

Federated learning (FL) may provide a solution to this problem.[4] The basic premise behind FL is simple: In the standard training of AI systems, data is collected and transferred onto a central server, where the AI system trains on the data. In contrast, in FL, data is not collected but remains in its original locations. Instead, each party receives the raw model, which is then trained on the dataset in situ. Upon completion of the training, the trained model is sent back, and is combined with other similarly trained models, into a single, integrated model. The result is that the integrated model has effectively been trained on all the datasets, but no data is transferred out from its original location.

Data is often held in ‘silos’.[5] A data silo is anything that holds data (e.g. smartphones, laptops, hospitals, banks, etc.) but accessing data held in silos is challenging. Sometimes regulation, like data protection legislation or IP law, prevents data from being shared. Alternatively, there may be a reluctance to share data, for instance, due to concerns about data confidentiality or data integrity. The intended purpose of FL is to ‘break open’ these data silos, by enabling the training of AI systems while preserving data privacy and confidentiality. If FL can fulfill this promise, this could bring significant benefits. By way of example, “healthcare providers could train algorithms to develop new drugs based on patient data, while maintaining privacy and patient confidentiality, or researchers in different countries could train algorithms without transmitting data across jurisdictions.”[6]

FL has generated significant interest amongst the computer science community, however, there is a dearth of writings on, and understanding of, FL among lawyers and legal academics.[7] This is a problem because the conceptualization of concepts like data and privacy may differ across disciplines[8] and to what extent FL can break open data silos created by regulation requires a legal analysis. This article aims to fill this lacuna by providing a comprehensive legal analysis of FL. This will be done by examining how the data protection principles – represented by the most stringent standards under the European Union’s General Data Protection Regulation (GDPR)[9] – applies to FL. The argument will made that from a legal perspective FL can indeed be an effective method to ensure compliance with data protection regulation.

Although the legal analysis in this article focuses on the GDPR, the significance of the analysis extends beyond the EU. EU regulation has proven influential beyond the EU,[10] and many data protection regimes are modelled on the GDPR.[11] Moreover, FL raises an important conceptual question about the relationship between data protection and the development of AI; that is, whether the training of AI systems on personal data is in itself an infringement of data protection rights, or whether there is such an infringement only because of some feature of how the training is conducted, e.g. that data is collected to a central server or access to the data is given to a third party. In the standard training of AI systems, this question will seldom arise as data needs to be collected for the AI systems to be trained, and many data protection regimes regulate the collection of data.[12] However, because in FL no data is collected, this issue is brought into sharp focus, and in the age of AI, this is a question every data protection regime will need to answer.[13]

This article suggests that the training of AI systems itself does not infringe data protection rights, provided that the data is kept secure from abuse (i.e. the data being used for purpose other than training AI). The argument is that using personal data to train AI systems does not reveal information about an individual, such information is only revealed when the AI system is applied to a particular case. This article will show that the GDPR can be interpreted in this way, and if this interpretation is followed, the GDPR can provide for the protection of personal data, without hindering the development of AI systems.[14] For the legal analysis of FL this means that the question of whether the training of an AI system through FL is GDPR compliant will largely depend only on one factor, namely whether the data is kept secure, rather than the host of factors, which is typically required to assess GDPR compliance in standard training of AI systems. Thus, FL should make it easier for AI developers to train models on personal data.

Despite FL being a potential boon to AI’s compliance with data protection regulation, this article will express doubt as to whether FL can make a significant contribution towards solving AI’s data problem. Although, FL may be an effective way to deal with data protection, data protection is only one among other obstacles to data sharing. For instance, IP law may prevent data from being shared, and FL does not directly impact the application of IP law. Moreover, it is unlikely that FL will be used sufficiently widely to make significantly more data available, than is currently the case. There is also a lack of legal clarity in relation to FL, and without legal clarity it is unlikely that FL will be commonly adopted. Furthermore, currently, FL is not used widely across different organizations.[15] This means that a lot of data will remain inaccessible. This is an area where regulators and policy makers may be able to make a positive contribution. This paper will suggest that, if regulators and policy makers decide to facilitate the use of FL, a possible tool is the creation of a FL regulatory regime, including an FL licensing regime, to facilitate data sharing across organizations.

This article will proceed as follows. First, an overview of FL will be provided. Second, this article will give a detailed analysis of how the GDPR applies to the training of AI systems. This analysis will take up considerable space, but it is crucial to understand how the GPPR applies to the training of AI systems as without such understanding, it is impossible to examine how the GDPR applies to FL. Third, this article will analyze to what extent FL can facilitate the sharing of non-personal data and examine the obstacles to FL being used more widely. Note that this article focuses on FL in relation to data protection regulation. Other issues, like IP law or antitrust law, will not be considered directly and are out of the scope for this article.

[1] Thilo Hagendorff & Katharina Wezel, 15 Challenges for AI: Or What AI (Currently) Can’t Co, 35 AI & Soc 355 (2020).

[2] Tal Roded & Peter Slattery, What Drives Progress in AI? Trends in Data, FutureTech (March 19, 2024), https://futuretech.mit.edu/news/what-drives-progress-in-ai-trends-in-data.

[3] Devika Rao, All-powerful, ever-pervasive AI is running out of internet, The Week (June 5, 2024), https://theweek.com/tech/ai-running-out-of-data; S.E. Whang, et al. Data collection and quality challenges in deep learning: a data-centric AI perspective, 32 VLDB J. 79 (2023).

[4] See Brendan McMahan & Daniel Ramage,, Federated Learning: Collaborative Machine Learning without Centralized Training Data, Google Research Blog (April 6, 2017), https://ai.googleblog.com/2017/04/federated-learning-collaborative.html.

[5] See Florian Gamper, Federated Learning: What Lawyers Need to Know, L.Gazette, (June 2024), https://lawgazette.com.sg/feature/federated-learning-what-lawyers-need-to-know/ (the Law Gazette is the official publication of the Law Sciety of Singapore).

[6] Id.

[7] However, there is some legal analysis of FL. See e.g. S. Rossello et al., Data Protection by Design in AI? The Case of Federated Learning, 116 Computerrecht (2021); Nguyen Truong et al. Privacy Preservation in Federated Learning: An Insightful Survey from the GDPR perspective, 110 Comput. Secur. J. 12402, 12414-18 (2021).

[8] The same claim could be made in relation to many other concepts, like transparency, bias, fairness, to mention just a few.

[9] Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation) 2016 O.J. (L119) [hereinafter GDPR].

[10] Anu Bradford, THE BRUSSELS EFFECT, 107 Nw. U. L. Rev. 1 (2012) (argues that EU regulation impact jurisdictions outside the EU).

[11] Graham Greenleaf, Now 157 Countries: Twelve Data Privacy Laws in 2021/22, 176 Privacy L. & Bus. Int’l Rep., 1, 1 (2022).

[12] GDPR, supra note 11, art. 4.1(2), (states that collecting is a form of processing, inter alia GDPR arts. 5 and 6 regulate processing).

[13] Just to clarify, the question is also relevant for jurisdictions which currently do not have a data protection regime but are considering creating such a regime.

[14] See Giovanni Sartor & Francesca Lagioia, The impact of the General Data Protection Regulation (GDPR) on artificial intelligence, European Parliamentary Rsch. Serv, 76 (June 2020) [hereinafter EPRS Study] (A study at the request of the Panel for the Future of Science and Technology (STOA) and managed by the Scientific Foresight Unit, within the Directorate-General for Parliamentary Research Services (EPRS) of the Secretariat of the European Parliament).

[15] Saikishore Kalloori & Abhishek Srivastava, Towards cross-silo federated learning for corporate organizations, 289 Knowledge-Based Sys. 1, (Apr. 8, 2024).

View the Entire Article

Share and Enjoy: