Alston & Bird Consumer Finance Blog

Privacy, Cyber & Data Strategy

NYDFS Issues Guidance on the Million Dollar Question—How to Conduct Cybersecurity Risk Assessments Under Part 500

By , , ,

Originally published September 21, 2026 on the Alston & Bird Privacy, Cyber & Data Strategy Blog.

On September 10, 2026, the New York State Department of Financial Services (“NYDFS”) published an Industry Letter (the “Letter”) to regulated entities (“Covered Entities”) addressing what is probably the most frequently asked Part 500 question: “How to Conduct and Use Risk Assessments Required by the DFS Cybersecurity Regulation.”

The risk assessment is, by all accounts, the core benchmark of Part 500. It permeates nearly every other substantive requirement of Part 500.

Let’s start with the basics. A risk assessment—by NYDFS’s definition—is “the process of identifying, estimating and prioritizing cybersecurity risks to organizational operations (including mission, functions, image and reputation), organizational assets, individuals, customers, consumers, other organizations and critical infrastructure resulting from the operation of an information system. Risk assessments incorporate threat and vulnerability analyses and consider mitigations provided by security controls planned or in place.” And how will it be used? NYDFS states it plainly: “The Department expects each Covered Entity to be able to demonstrate how its Risk Assessment informed cybersecurity controls, compensating controls, and risk acceptance decisions.”

This is, as with all things, easier said than done. And while the Letter states that it does not create new legal obligations, it does provide specificity surrounding NYDFS’s expectations, flags common issues, and identifies best practices. Regardless of whether the Letter does or does not identify net new obligations, it is safe to assume that NYDFS will rely heavily on this interpretation in future examinations and investigations.

What NYDFS Says Gets Missed

Leveraging its historical examinations and investigations, NYDFS has identified certain “gaps” in risk assessments that, in its view, contribute to a “deficient” program:

  • Incomplete asset scope and visibility, including outdated or incomplete asset inventories; failing to identify where NPI resides or flows; and omitting critical business processes, Third-Party Service Providers,11 cloud environments, or other external dependencies.
  • Weak or inconsistent methodologies, including failing to consistently identify, analyze, prioritize, and document cybersecurity risks; evaluate the effectiveness of existing controls; or distinguish between inherent and residual risk.
  • Failure to account for evolving and interconnected risks, including emerging technologies, changes in the threat landscape, interdependencies, concentration risk, and single points of failure that could materially affect the Covered Entity’s operations.
  • Insufficient governance and risk treatment, including failing to assign ownership, document risk response decisions, integrate Risk Assessment results into enterprise governance, or update Risk assessments following material changes to the business, technology, or threat environment.
  • Failure to account for or inform the cybersecurity program, resulting in policies, controls, and resource decisions that are not demonstrably based on the Covered Entity’s identified cyber risks.

The Risk Assessment NYDFS Says It Wants

The Letter hones in on five core areas of the risk assessment that are integral, in its assessment, to a successful, mature program: (1) Governance and Oversight; (2) Defined and Repeatable Methodology; (3) Scope and Coverage; (4) Documentation and Traceability; and (5) Integration and Updates to Risk Assessments.

Governance and Oversight

Second to maybe only AI, the Letter starts with recommendations regarding the top buzzword of 2026: Governance. The Department’s guidance here is simple: risk management is a collective responsibility. While the risk assessment process must be overseen by the top security official, it should be cross-functional in nature to ensure there is a comprehensive understanding of risk enterprise wide.

Though not mandatory, the Letter also strongly encourages reporting to senior management and using the risk assessment to help “make informed decisions regarding resource allocation, cybersecurity investments, control selection, and risk acceptance.”

Defined and Repeatable Methodology

A recurring theme throughout the Letter is the need for a documented, repeatable methodology. Process, process, process. NYDFS expects covered entities to establish a structured mechanism for identifying, analyzing, and prioritizing risks using consistent criteria. Among the inputs to be considered, NYDFS identifies the following:

  • Threat intelligence, incident trend analysis, vulnerability scans, and penetration testing;
  • Findings from audits or prior risk assessments;
  • Technical threats (whether internal or external), including malicious actors, system misconfigurations, insider misuse, and process failures;
  • Administrative controls;
  • Natural disasters that could cause power outages, disrupt data centers, and impair system availability;
  • Third party dependencies that could be exploited by bad actors.

NYDFS emphasizes the need to provide reasonable estimates of the likelihood and impact of risks and to apply consistent rating criteria to ensure an apples-to-apples comparison over time. And as for what “impacts” should be considered, NYDFS identifies a broad range, including:

  • Exposure of sensitive personal information;
  • Financial losses;
  • Business interruptions;
  • Regulatory or legal liability;
  • Reputational damage; and
  • Costs of recovery and replacement.

Although NYDFS does not mandate any particular framework, it does strongly suggest that the use of an existing, recognized framework (for instance, NIST CSF 2.0, the Cyber Risk Institute Profile, and ISO 27005) leads to a more consistent and repeatable methodology.

NYDFS closes the section with advice on how to bring a program to the next level of maturity—applying a consistent risk criteria beyond just the organization itself, but also to its third-party risk management, IT operations, and BCDR planning.

Scope and Coverage

Once Covered Entities determine the method of conducting their risk assessment, the next step is to consider its scope. The Department recommends focusing on any internal and external factors that may materially affect cyber risk. The Letter specifically highlights four areas of emphasis: all assets, emerging risks, third-party risk, and concentration risk.

  • All Assets. NYDFS warns that the failure to account for assets in the risk assessment process in turn leads to an incomplete view of risk and, at times, can result in material gaps in the cybersecurity program.
  • Emerging Risks. Consistent with NYDFS’s recent guidance regarding the heightened cybersecurity threat landscape, Covered Entities should be cognizant of advancements in artificial intelligence, evolving ransomware techniques, supply chain threats, and geopolitical developments that may affect the cyber threat landscape.
  • Third-Party Risk. Consistent with its prior guidance regarding the risks associated with Third-Party Service Providers, NYDFS recommends that Covered Entities evaluate third-party risk based on factors such as connectivity, data access, operational dependency and the sensitivity of the information involved.
  • Concentration Risk. The Department encourages thinking holistically. For example, a company may use the same cloud provider, software platform, identity provider, or managed service provider across multiple business functions. Each relationship may appear low risk on its own, but if that shared provider suffers a cyberattack or outage, it could disrupt numerous critical systems at the same time.

Documentation and Traceability

The Letters places substantial emphasis on documentation. Covered entities should be able to demonstrate not only that risks were identified, but also how those risks informed cybersecurity controls, compensating controls, remediation activities, and risk acceptance decisions.

NYDFS recommends maintaining documentation that links identified risks to specific mitigation measures and preserving the rationale supporting management decisions. The Department also encourages the use of a risk register or similar tracking mechanism to monitor remediation efforts and changes in residual risk over time.

From an examination perspective, this portion of the guidance suggests that regulators will increasingly expect organizations to show evidence of decision-making rather than simply produce policies or assessment reports.

Integration and Updates to Risk Assessments

NYDFS reiterates that risk assessments cannot be treated as static, annual compliance exercises. While annual reviews remain mandatory, covered entities must also update assessments whenever material changes occur. Examples identified by the Department include mergers and acquisitions, major system migrations, significant outsourcing arrangements, adoption of emerging technologies, evolving threat actor tactics, and major geopolitical events.

The Letter strongly favors continuous or regularly refreshed assessment processes that evolve alongside changes in technology, business operations, and the threat environment.

But What Does This Mean?

For regulated entities, the most immediate takeaway is that risk assessments should be reviewed against the Letter to determine whether they:

  • Capture all relevant assets and data flows;
  • Address emerging and interconnected cyber risks;
  • Include robust governance and stakeholder involvement;
  • Create a clear linkage between risks and cybersecurity controls; and
  • Support ongoing updates in response to changes in business operations and technology.

NYDFS’s latest guidance provides a detailed picture of what the Department considers a mature, defensible cybersecurity risk assessment program. While framed as interpretive guidance, it is likely to serve as a benchmark in future examinations, investigations, and enforcement actions. Organizations subject to Part 500 should view the publication as both a compliance resource and a preview of the standards against which their risk assessments may increasingly be judged.

Privacy, Cyber & Data Strategy Advisory | Five Risks GCs Should Know About Frontier AI and Cybersecurity

By , ,

Originally published June 1, 2026 on the on the Alston & Bird website.

Executive Summary

Our Privacy, Cyber & Data Strategy Group explores how frontier AI models are changing the cybersecurity landscape and what general counsel need to know about the risks and challenges for organizations.

  • Frontier AI models are increasing the speed and sophistication of cyber threats
  • Organizations may have less time to identify and respond to vulnerabilities
  • General counsel should assess whether existing cybersecurity practices remain effective in light of evolving threats

The rapid advancement of frontier artificial intelligence (AI) models, such as Anthropic’s Claude Mythos, OpenAI’s GPT-5.5 Cyber, and Google Gemini, has introduced a shift in the cybersecurity landscape that demands the attention of every general counsel. These models can identify software vulnerabilities and methods to exploit them at machine speed, find security vulnerabilities that traditional security tools cannot detect, and present ways of chaining together multiple vulnerabilities, including low-severity ones, to exploit and compromise systems.

Together, these capabilities are expected to alter the threat environment in which companies operate. General counsel should consider five critical areas as they evaluate the risks posed by frontier AI models.

1. The Genie Is Out of the Bottle

The question is no longer whether frontier AI models will be used to discover and exploit software vulnerabilities at scale, but when—and the window between “when” and “now” is closing rapidly. Researchers and threat actors have demonstrated that large language models and other frontier systems can analyze codebases, identify zero-day vulnerabilities, and generate working exploit code with striking efficiency. What once required weeks or months of skilled human effort can now be accomplished in hours or minutes.

For general counsel, the practical implication is straightforward: the volume, speed, and sophistication of cyberattacks targeting your organization are poised to increase dramatically. The historical pace at which vulnerabilities were discovered and exploited provided organizations with a reasonable, if imperfect, remediation window. That window is closing. Legal and compliance teams must internalize this shift and begin planning accordingly, because the liability exposure that accompanies a failure to adapt will be significant.

2. Expect Increased Exposure Across the Enterprise

One of the most consequential effects of AI-accelerated vulnerability discovery is the compression of the timeline between identification and exploitation. Threat actors have long exploited zero-day vulnerabilities. For the broader universe of vulnerabilities, however, organizations historically relied on a period of relative grace, during which patches could be developed, tested, and deployed before widespread exploitation occurred. Frontier models threaten to collapse that timeline to zero far more routinely, making what was once the exception increasingly the norm.

This creates two distinct and compounding problems. First, the sheer volume of newly discovered vulnerabilities may overwhelm even well-resourced IT and security teams. Prioritization, already a challenge, becomes even more difficult when the queue of critical vulnerabilities grows faster than the organization’s capacity to remediate them.

Second, and perhaps more concerning for general counsel, many systems on which organizations rely are maintained by third-party vendors. Internal teams do not have access to the underlying source code and cannot independently develop or deploy patches. Organizations are therefore dependent on vendors’ responsiveness. If a critical vendor is slow to issue a patch, or overwhelmed by the volume of vulnerabilities in its products, the organization remains exposed regardless of its internal security posture. This reliance on third parties to remediate vulnerabilities in software the organization neither controls nor can inspect presents a legal and operational risk that general counsel must squarely confront.

3. A Documented Cyber Risk Strategy Will Help Mitigate Risk

In the face of these evolving threats, one of the most important steps a general counsel can take is to work closely with the chief information security officer (CISO) and broader information security team to develop and document a comprehensive cyber risk strategy. This is particularly important as regulators increasingly take note of the risks posed by frontier AI models.

A sound strategy should begin with a structured framework for prioritizing identified vulnerabilities based on factors such as exploitability, potential business impact, availability of patches or mitigations, and the criticality of affected systems. Build an approach that is tailored to your specific risk profile. The strategy should also include clear protocols for communicating cyber risk to senior executives and the board of directors. Boards are expected to exercise informed oversight of cybersecurity matters, and general counsel play a critical role in ensuring that information flowing to the board is accurate, timely, and actionable. A well-documented communication cadence covering vulnerability posture, emerging threats, and remediation progress helps enable informed decision-making and demonstrates that the organization is taking these risks seriously.

A documented strategy also helps establish a defensible position in the event of a breach. If an organization suffers a security incident before it has been able to patch a known vulnerability, regulators, plaintiffs’ counsel, and courts will examine the steps taken upon learning of the vulnerability and whether those steps were reasonable. An organization that can demonstrate a risk-based prioritization framework, supported by contemporaneous documentation showing how decisions were made and resources allocated, is in a stronger position than one that cannot.

Regulatory expectations regarding what constitutes “reasonable security” may continue to rise in response to the AI-driven threat landscape. Regulators have signaled that the standard of care is not static and that organizations are expected to adapt their security practices as threats evolve. For example, financial regulators in the United States and the United Kingdom have noted that companies deploying AI security tools to defend against threats may be better able to mitigate the risks associated with frontier AI models. The Office of the Comptroller of the Currency (OCC), in its Spring 2026 Semiannual Risk Perspective, highlighted the role of AI in defending against threats and supporting risk management and enhanced threat and vulnerability monitoring processes. Similarly, the New York Department of Financial Services (NYDFS), in guidance released on May 21, 2026, stated that frontier AI models have materially changed cybersecurity risks and may warrant stronger defensive measures. A documented strategy that accounts for these risks helps position organizations to meet that standard.

4. Understand Which Critical Vendors You Rely On

Third-party dependency is a significant risk amplifier in this environment. General counsel should work with procurement, IT, and information security teams to identify vendors on which the organization is most critically dependent, particularly those handling sensitive data, supporting critical business functions, or deeply embedded in the organization’s infrastructure.

Once vendors are identified, the next step is proactive engagement. Organizations should engage with critical vendors now to understand how they are addressing threats posed by frontier AI models and similar tools. Key questions include:

  • What is the vendor’s current patch development and deployment timeline?
  • Has the vendor invested in AI-driven defensive capabilities to accelerate its own vulnerability detection and remediation?
  • Does the vendor have a documented incident response plan that accounts for AI-accelerated attacks?

These conversations should occur before an incident. General counsel should also review vendor agreements to assess whether current terms adequately address evolving risks, including indemnification provisions, service level agreements related to security patching, and notification obligations. Where gaps exist, renegotiation should be prioritized.

5. Consider Engaging a Vendor to Conduct a Readiness Assessment

General counsel should also consider engaging a cybersecurity firm to conduct a readiness assessment focused on preparedness for AI-driven threats. A number of cybersecurity consulting and incident response firms now offer assessment services designed to evaluate an organization’s ability to detect, respond to, and recover from attacks enabled by frontier AI models.

A readiness assessment may include:

  • Evaluation of vulnerability management processes and their ability to scale.
  • Examination of monitoring and detection capabilities for AI-generated attack techniques.
  • Review of incident response preparedness, including tabletop exercises incorporating AI-accelerated attack scenarios.
  • Evaluation of third-party risk management programs, including visibility into vendors’ security posture.

Engaging a reputable firm and acting on its recommendations can strengthen the organization’s ability to demonstrate that it has taken reasonable steps to prepare for foreseeable threats.

Conclusion

The intersection of frontier AI models and cybersecurity is not a future concern. It is a present and rapidly evolving reality. General counsel are well positioned to drive the coordination, strategic planning, and vendor engagement required to respond. While not exhaustive, the five steps outlined above offer a practical starting point for addressing these risks now.

Privacy, Cyber & Data Strategy Advisory | How AI Is Changing the Incident Response Landscape: What GCs Need to Know

By , ,

Originally published January 6, 2026 on the Alston & Bird website.

Executive Summary

Our Privacy, Cyber & Data Strategy Team examines the profound implications of the evolution of AI-driven cyberattacks and offers practical steps general counsel can take to proactively defend against them.

  • Vibe hacking—AI’s ability to work autonomously—eliminates threat actors’ need for technical expertise or large teams to conduct complex attacks
  • Game-changing polymorphic malware and “just-in-time” AI-driven code regeneration can defeat traditional defenses
  • General counsel can update their companies’ own AI systems and ensure third-party AI vendors have the latest security

The cyber-threat landscape has always evolved rapidly, but the emergence and weaponization of artificial intelligence (AI)—particularly generative AI (GenAI)—by threat actors represents a seismic shift that cannot be ignored. Just one year ago, we wrote about the early stages of adversaries using AI to automate and customize cyberattacks, primarily to improve phishing campaigns, develop deepfakes, and refine their tactics, techniques, and procedures. Over the past year, threat actors have significantly escalated their use of AI, moving from “vibe hacking”—the use of agentic AI systems that can reason, plan, and act autonomously as both technical consultants and active operators of cyberattacks—to unprecedented integration and autonomy of AI throughout the attack life cycle.

If that were not enough, we are now starting to see polymorphic malware and “just-in-time” AI-driven code regeneration, which are true game changers. Polymorphic malware powered by AI can continuously rewrite its own code in real time, defeating traditional signature-based detection and even heuristic analysis. Compounding these risks are cyberattacks targeting AI systems themselves, such as prompt injection attacks that manipulate the reasoning layer of AI models and often leave no meaningful forensic trail. This creates significant limitations for conventional forensic tools and logging frameworks, which were never designed to capture the internal logic of autonomous AI agents.

The Evolution of AI in Cyberattacks: From Phishing Emails and Deepfakes to Fully Automated AI-Powered Cyberattacks

Initially, threat actors used AI to enhance phishing campaigns—improving grammar, tone, and personalization to increase success rates. Attackers soon leveraged AI to create convincing deepfakes for social engineering and fraud. By mid-2025, vibe hacking emerged as hackers began to use and code agentic AI systems not just to be assistants in an attack but as autonomous operators capable of executing complex, multistep cyberattacks. Today, organizations face fully automated AI-powered attacks with minimal human oversight.

This shift is driven by several factors:

  • An evolution of AI to be able to execute nearly every stage of an attack.
  • The maturing underground marketplace for illicit AI tools.
  • AI’s ability to complete cyberattacks faster than before.
  • Elimination of the need for hackers with deep technical expertise or large teams with specific expertise to conduct complex cyberattacks.

Vibe hacking is threat actors’ use of agentic AI systems as technical consultants and active operators of cyberattacks. In practice, AI agents like Claude Code are no longer just assisting attackers; they are executing multistep operations independently, from reconnaissance and credential harvesting to data exfiltration and ransom note generation. This evolution lowers the barrier to entry for sophisticated cybercrime and enables lone actors to conduct campaigns that previously required coordinated teams.

Cybercriminals took this to the next level in the last quarter of 2025. Highly sophisticated threat actors, including a Chinese state-sponsored group (designated by Anthropic as GTG-1002), demonstrated “unprecedented integration and autonomy of AI throughout the attack lifecycle.” GTG-1002 coordinated targeted attacks against approximately 30 entities, with several confirmed compromises.

Claude Code was manipulated to autonomously execute 80%–90% of an attack—reconnaissance, vulnerability discovery, exploitation, lateral movement, credential harvesting, data analysis, and exfiltration. Humans primarily assumed a strategic oversight role, initiating the campaign and intervening at critical decision points. Anthropic indicated this marked the first documented case of agentic AI successfully obtaining access to confirmed high-value targets for intelligence collection.

While these attacks were highly sophisticated, limitations remain. In the Anthropic-reported attacks, AI tools occasionally hallucinated data, misidentified credentials, and flagged publicly available information as sensitive. These errors, however, can be corrected with minimal human oversight, rapidly shrinking the limitations to fully autonomous attacks.

Accelerated speed of AI-powered cyberattacks

AI has dramatically shortened the time required to execute a cyberattack from start to finish. In fact, a defining characteristic of AI-powered attacks is the ability to gather and analyze data efficiently. AI tools not only speed up the research phase of a cyberattack but also improve the accuracy and completeness of a threat actor’s analysis of data.

Anthropic’s November 2025 report illustrates how AI compresses every stage of the attack life cycle:

  • Reconnaissance. AI agents can automate reconnaissance tasks, scanning thousands of endpoints, cataloging systems, mapping infrastructure, identifying exposed services, and analyzing authentication mechanisms quickly. Tasks that once required weeks of manual effort can now be completed in hours.
  • Vulnerability Discovery. AI systems can autonomously test for misconfigurations, weak credentials, and exploitable flaws using adaptive scripts that evolve based on real-time feedback.
  • Exploitation and Lateral Movement. Agentic AI can sequentially leverage multiple vulnerabilities without human intervention, moving across networks and escalating privileges with minimal oversight.
  • Credential Harvesting and Data Analysis. AI accelerates credential cracking and can parse massive datasets for sensitive information at machine speed, enabling attackers to identify and prioritize high-value assets instantly.
  • Exfiltration and Monetization. AI orchestrates stealthy data exfiltration while simultaneously generating ransom notes or negotiating scripts tailored to victims.

This level of automation is not theoretical—it has already been observed in real-world campaigns. The ability to compress what was once a multiweek operation into a matter of hours or days is a paradigm shift in cyber-risk.

Lowering the barrier to entry

AI has erased the need for hackers to have extensive technical expertise. Threat actors who previously lacked the skills to execute complex attacks can now simulate professional competence through AI assistance. For example, North Korea’s IT remote worker scheme has been transformed by AI, enabling operatives to pass interviews, maintain engineering roles, and deliver work product without formal training. Beyond nation-state actors, AI empowers individual cybercriminals to accomplish what once required entire teams. Developing ransomware, identifying targets, and making strategic and tactical decisions around exploitation and monetization of stolen data can now be performed with the assistance of AI and a single individual overseeing the operation.

AI-Powered Polymorphic Malware and Code Regeneration

In its November 2025 AI Threat Tracker report, Google’s Threat Intelligence Group (GTIG) highlighted a shift from threat actors leveraging AI just for productivity gain to threat actors deploying novel AI-enabled malware in active operations. This can be seen in the growing use of just-in-time (JIT) code regeneration, allowing attackers to dynamically rewrite malicious code during execution, making detection and static analysis extremely difficult. GTIG emphasized that this capability enables malware to adapt to defensive measures in real time, reducing the effectiveness of traditional signature-based detection.

While GTIG focused on JIT regeneration, other sources, including Anthropic and independent threat researchers, have documented the rise of AI-powered polymorphic malware—malicious code that continuously mutates its structure to evade detection. Emerging strains such as PROMPTFLUX demonstrate how attackers use large language models (LLMs) to adjust code in malware so it can evolve dynamically and remain stealthy almost indefinitely.

The implications are profound: AI-driven polymorphism and JIT regeneration reduce the cost and complexity of maintaining stealth, making advanced malware accessible to less-skilled actors. Combined with AI’s ability to orchestrate reconnaissance, exploitation, and exfiltration, this trend alters the landscape for enterprise security.

Cyberattacks on AI Systems and Investigation Gaps

Cyberattacks against AI systems are introducing new challenges for forensic investigations and for understanding what happened and why. Even with the right security such as endpoint detection and response software and robust traditional monitoring and logging in place, there may be significant gaps when investigating AI-driven attacks.

One of the most concerning examples is a prompt injection attack, which targets AI systems by embedding malicious instructions inside what appears to be normal input. For instance, an attacker might hide a command such as “delete all logs” within a seemingly benign request. Because LLMs cannot reliably distinguish between trusted commands and untrusted data, the AI may execute these instructions without question. This creates a fundamental problem both for securing the AI system itself and for forensically investigating these attacks.

Conventional investigations rely heavily on system-level logs—detailed records that explain what happened within an operating system or application—that allow investigators to reconstruct the timeline of an attack, identify actions taken, and determine which systems were compromised. However, prompt injection attacks occur inside the AI’s reasoning layer, not at the operating system level. If the AI is instructed to erase or alter logs, investigators may only see the outcome (e.g., data exfiltration or deletion) without any record of the causal chain or how the outcome occurred.

In other words, the “why” behind the action is missing because the attack exploits the AI’s cognitive process rather than the system’s technical process. Forensic teams cannot rely on the old playbook—they likely need new methods such as monitoring AI interactions, auditing prompts, and implementing specialized AI audit trails to reconstruct the reasoning chain.

Practical Tips and Actionable Steps

As AI-driven cyber-threats accelerate, general counsel (GCs) play a critical role in shaping governance, risk management, and legal response strategies.

  1. Update Incident Response Procedures for AI-Powered Cyberattacks. AI-driven attacks occur at unprecedented speed and often outpace traditional response timelines. Companies should consider updating incident response procedures to explicitly address these threats. Organizations should also consider incorporating scenarios involving AI-powered attacks, such as polymorphic malware and prompt injection, into tabletop exercises to test readiness and identify gaps. These exercises provide valuable insights into how an AI-driven incident might unfold and help better prepare the organization to respond to the unique nature of AI-driven attacks.
  2. Investigate AI-Powered Cyberattacks—Protect Privilege and Ensure Vendor Expertise. AI-powered attacks introduce new types of evidence—such as prompt logs, model outputs, and reasoning steps—that traditional forensic processes do not capture. GCs should structure investigations to preserve attorney-client privilege for AI-related forensic evidence. Given the new types of forensic evidence, companies may also want to verify that their preferred third-party forensic firms maintain the necessary expertise.
  3. Audit AI Inputs. Where feasible and appropriate, organizations should consider regularly auditing AI inputs to fine-tune the AI system to detect and block malicious or misleading prompts before they trigger harmful actions. Auditing reduces the risk of unauthorized activity and provides transparency into how AI systems interpret and act on sensitive information. Reviewing prompts and outputs enables companies to identify misuse patterns; it also strengthens controls and ensures accountability in AI-driven decision-making.
  4. Revisit Vendor Management and Contracts. AI introduces risks that traditional vendor agreements do not fully address. Companies should consider contractual provisions that require vendors to monitor AI systems for misuse, such as prompt injection attacks, and maintain detailed audit trails of prompts and outputs for forensic investigations. Companies should consider including provisions that mandate compliance with emerging AI regulations in vendor agreements to ensure vendors meet evolving legal and security standards.
  5. Review Governance and Oversight. GCs can be an advocate for an AI risk governance framework that aligns with regulatory expectations and industry standards. Board-level reporting should include AI-specific threat trends and mitigation strategies. This helps with leadership visibility and accountability for AI-related risks.
  6. Monitor Regulatory and Liability Developments. Stay ahead of emerging AI regulations and assess potential liability exposure for AI misuse or compromised AI systems. This will allow GCs to advise leadership on compliance obligations and risk mitigation strategies.

Ransomware Fusion Center

Stay ahead of evolving ransomware threats with Alston & Bird’s Ransomware Fusion Center. Our Privacy, Cyber & Data Strategy Team offers comprehensive resources and expert guidance to help your organization prepare for and respond to ransomware incidents. Visit Alston & Bird’s Ransomware Fusion Center to learn more and access our tools.

If you have any questions, or would like additional information, please contact one of the attorneys on our Privacy, Cyber & Data Strategy team.

You can subscribe to future advisories and other Alston & Bird publications by completing our publications subscription form.

NYDFS Issues Guidance on Managing Risks Related to Third-Party Service Providers

By ,

On October 21, 2025, the New York Department of Financial Services (“NYDFS”) published an Industry Letter (the “Letter”) outlining guidance on managing risks related to third-party service providers (“TPSPs”). NYDFS recognizes that as covered entities become more reliant on TPSPs, managing TPSPs “remains a crucial element of a Covered Entity’s cybersecurity program.” The Letter outlines the actions and advice Covered Entities should take while progressing through the lifecycle of a TPSP relationship: (1) Identification, Due Diligence, and Selection; (2) Contracting; (3) Ongoing Monitoring and Oversight; and (4) Termination. While the Letter expressly states it does not impose new requirements or obligations on Covered Entities, but rather, intended to clarify Part 500 (specifically 500.11) and recommend best practices, the prescriptive guidance may in practice be considered the operative benchmark for certain (or many) Covered Entities.

Identification, Due Diligence, and Selection

Due to the increased risks associated with TPSP relationships, Covered Entities may wish to exercise caution and diligence before entering into any arrangement with a TPSP. Accordingly, the Letter outlines a non-exhaustive list of considerations for conducting due diligence on TPSPs. A few of those considerations include:

  • The type and extent of the TPSPs access to information systems and nonpublic information (“NPI”);
  • The TPSPs reputation within the industry, including its cybersecurity history and financial stability;
  • The controls the TPSP has implemented for its own systems and data, particularly if the Covered Entity’s systems are not fully segregated;
  • Whether the TPSP undergoes external audits and independent assessments;
  • The TPSPs practices for selecting, monitoring, and contracting with downstream service providers; and
  • Whether the TPSP, its affiliates, or vendors operate in or from jurisdictions considered high-risk due to geopolitical, legal, socio-economic, operational, or regulatory factors.

In addition to the above considerations, the Letter emphasizes that Covered Entities should consider how to best obtain, review, and validate information provided by prospective TPSPs. Although standardized questionnaires may facilitate the process of gathering the required information from the TPSPs, Covered Entities must ensure that those questionnaires are interpreted by qualified personnel to allow for proper risk-informed decisions to be made. In other words, vendor due diligence questionnaires are not a “check-the-box” requirement; the completed questionnaires must be carefully evaluated by qualified personnel and actioned appropriately. Additionally, if there are limited vendor options, Covered Entities should make risk-informed decisions, document the relevant risks, and take steps to implement compensating controls.

Contracting

Covered Entities that utilize TPSPs are required to develop and implement written policies and procedures that address due diligence and contractual protections. In the Letter, NYDFS provides a few examples of “baseline contract[al] provisions” that Covered Entities should consider incorporating into agreements with a TPSP. Some of the provisions include:

  • Develop and implement policies and procedures addressing access controls;
  • Develop and implement policies and procedures addressing encryption in transit and at rest;
  • Provide immediate or timely notice to the Covered Entity upon occurrence of a Cybersecurity Event directly impacting the Covered Entity’s information;
  • Require TPSPs to disclose where data may be stored, processed, or accessed; and
  • Require TPSPs to disclose the use of subcontractors and allow the Covered Entity to reject the use of certain subcontractors, which essentially allows Covered Entities the ability to control the use of Fourth Parties, somewhat akin to GDPR.

In addition to the above provisions, the Letter reinforces similar guidance that NYDFS provided in 2024, which we previously covered, in regard to inserting provisions in TPSP agreements that relate to the acceptable use of Artificial Intelligence (“AI”) products.

NYDFS clarified that the list provided in the Letter is neither exhaustive nor appropriate in all situations, but Covered Entities should continue to seek “reasonable protections, such as breach notification clauses, data use, and assurances regarding access controls and data handling.” Further, Covered entities should develop medium- to- long-term strategies to reduce its overall dependency on TPSPs.

Ongoing Monitoring and Oversight

A Covered Entity that utilizes a TPSP must have policies in place addressing the periodic assessment of TPSPs based on the risk that each “TPSP presents and the continued adequacy of [the TPSPs] cybersecurity practices.” The assessments conducted by Covered Entities may include obtaining security attestations from the TPSPs (e.g., SOC2, ISO 27001) and requiring penetration testing summaries, policy updates, evidence of security awareness training and proof of compliance audits.

In addition to the periodic assessments, Covered Entities should request updates on a TPSPs vulnerability management, assess patching practices, and confirm remediation of previously identified deficiencies. Although it may be an extensive exercise for Covered Entities, the Letter indicates that Covered Entities should document material or unresolved risks identified and escalate the risks as appropriate.

Termination

When a Covered Entity terminates its relationship with a TPSP, there are actions the Covered Entity should take to mitigate any potential risks from arising. Some of the actions the Letter outlines are prescriptive, including:

  • Revoking identity federation tools, API integrations, and external storage access;
  • Requiring certification of destruction of NPI, secure return of data, or migration of data to another TPSP or internal environment;
  • Confirming that any remaining snapshots, backups, or cached datasets are deleted and access to any shared resources is revoked;
  • Giving special attention to residual or unmonitored access points that fall outside routine access provisioning systems; and
  • Engaging keys take holders, including IT, legal, compliance, procurement, and business units to identify strategies to mitigate potential risks when planning to terminate.

Notably, in addition to the above actions a Covered Entity should take during the termination period, Covered Entities should ensure the offboarding process is properly documented and all relevant audit logs are retained to support accountability and future verifications.

NYDFS made it clear that it will continue to “consider the absence of appropriate TPSP risk management practices by Covered Entities in its examinations, investigations, and enforcement actions.” As such, although the Letter does not formally impose any new requirements for Covered Entities, Covered Entities are strongly encouraged to review the Letter and implement the practices identified by NYDFS to strengthen their cybersecurity posture.

This post was originally published on Alston & Bird’s Privacy, Cyber & Data Strategy Blog on October 27, 2025.

Privacy, Cyber & Data Strategy / White Collar, Government & Internal Investigations Advisory | GENIUS Act Establishes Federal Regulatory Oversight of Global Stablecoin Industry

By , ,

Executive Summary
8 Minute Read

Our Privacy, Cyber & Data Strategy and White Collar, Government & Internal Investigations Teams examine how the GENIUS Act’s framework for stablecoin issuers will impact the cryptocurrency sector.

  • The Act restricts the issuance of payment stablecoins within the United States to “permitted payment stablecoin issuers” (PPSIs)
  • PPSIs must maintain reserves of high-quality, liquid assets that fully back their outstanding stablecoins on at least a one-to-one basis
  • Regulatory oversight is divided between federal and state authorities, with joint oversight applying when state issuers exceed certain thresholds or opt into federal frameworks

___________________________________________________

On July 17, 2025, during “Crypto Week,” the U.S. House of Representatives passed the landmark Guiding and Establishing National Innovation for U.S. Stablecoins Act (GENIUS Act). Signed into law by President Donald Trump the next day, the GENIUS Act establishes a comprehensive federal framework for the issuance of payment stablecoins, regulation of stablecoin issuers, and both federal and state oversight for stablecoin authorization, audits, and other obligations. Domestic and foreign issuers in the more than $250 billion stablecoin market now have a clear path to securing and maintaining regulatory compliance in the United States.

Demonstrating rare cross-aisle cooperation and a shared interest in modernizing financial regulations to match emerging blockchain and artificial intelligence (AI) technologies, the Act garnered 308 affirmative votes in the House and 68 in the Senate, surpassing the upper chamber’s filibuster threshold. The GENIUS Act addresses Trump’s key campaign and policy promise to bring clarity and control to the digital asset market.

Key Provisions of the GENIUS Act

Effective date

The GENIUS Act takes effect on the earlier of (1) January 18, 2027 (18 months after the date the Act is enacted into law); or (2) 120 days after the primary federal regulators responsible for stablecoins issue their final regulations to implement the Act.

Authorized issuance of stablecoins only

The Act restricts the issuance of payment stablecoins within the United States to only those entities that qualify as “permitted payment stablecoin issuers” (PPSIs). PPSIs must be either U.S.-based issuers authorized under the Act or foreign issuers that are registered and operate under a regulatory framework deemed comparable to the Act by U.S. authorities and are subject to supervision by the Office of the Comptroller of the Currency (OCC).

A domestic PPSI must meet the requirements of one of three main categories: (1) subsidiary of an insured depository institution that has received approval to issue payment stablecoins under Section 5 of the Act; (2) federal qualified payment stablecoin issuers, which encompass nonbank entities (excluding state-qualified issuers) approved by the OCC, uninsured national banks chartered and approved by the OCC, or a foreign bank that does business outside the United States and has opened one or more federally licensed branches or offices in a U.S. state (“federal branch”), approved by the OCC; or (3) state-qualified payment stablecoin issuers, which are entities legally established under state law and approved by a state payment stablecoin regulator, provided they are not an uninsured national bank, federal branch, insured depository institution, or subsidiary of any such entities.

Requirements for issuing stablecoins

PPSIs must maintain reserves that fully back their outstanding stablecoins on at least a one-to-one basis. These reserves must consist of high-quality, liquid assets such as U.S. coins and currency or credit with a Federal Reserve Bank, demand deposits at insured depository institutions, short-term U.S. Treasury securities, and other monetary securities described in Section 4(a)(1) of the GENIUS Act. Any PPSI must publicly disclose its redemption policies and publish monthly reports detailing the composition, average maturity, and custody location of its reserves. A PPSI’s CEO and CFO must certify the accuracy of those monthly reports, and the Act makes knowingly false certifications punishable by up to 10 or 20 years’ imprisonment under 18 U.S.C. § 1350. To ensure reserve quality and transparency, PPSIs are prohibited from pledging, rehypothecating, or reusing reserves except under limited conditions, such as meeting margin obligations for investments in permitted reserves or creating liquidity to redeem payment stablecoins.

Mitigating money laundering and illicit financing risk

The GENIUS Act designates permitted payment stablecoin issuers as “financial institutions” under the Bank Secrecy Act (BSA), requiring them to implement robust compliance programs to prevent money laundering, terrorist financing, sanctions evasion, and other illicit activity. PPSIs must annually certify that they have implemented an effective BSA/AML compliance program. False certifications are punishable by up to five years’ imprisonment. To ensure regulatory parity, the Act’s registration and inspection requirements for foreign issuers effectively subjects them to similar compliance standards when accessing the U.S. market. Issuers must also be technologically capable of assisting with asset freezes, seizures, and turnovers pursuant to lawful orders. The Act further strengthens enforcement by requiring both U.S. and foreign issuers to (1) maintain the technical ability to comply with such orders; and (2) comply with them. Foreign issuers that fail to do so may be designated “noncompliant” by the Treasury, triggering a ban on secondary trading of their stablecoins after 30 days. Violations of that ban carry steep penalties—up to $100,000 per day for digital asset service providers and $1 million per day for foreign issuers.

Regulatory oversight

Regulatory oversight is divided between federal and state authorities, with federal regulators overseeing federally chartered or bank-affiliated issuers, state regulators supervising state-chartered issuers, and joint oversight applying when state issuers exceed certain thresholds or opt into federal frameworks. Regulators are responsible for licensing, examining, and supervising PPSIs to ensure compliance with the Act’s requirements, including reserve backing, redemption policies, and risk management standards.

PPSIs with more than $50 billion in consolidated total outstanding issuance that are not subject to the reporting requirements of the Securities Exchange Act of 1934 are required to prepare an annual financial statement in accordance with generally accepted accounting principles (GAAP) and must disclose any “related party transactions,” as defined under GAAP. A registered public accounting firm must audit the annual financial statement, and the audit must comply with all applicable standards set by the Public Company Accounting Oversight Board. These audited financial statements must also be made publicly available on the PPSI’s website and submitted annually to the PPSI’s primary federal payment stablecoin regulator.

Civil and criminal penalties

Additional civil and criminal penalties are set out throughout the Act. Notably, entities other than PPSIs that issue payment stablecoins in the United States without proper approval may face civil penalties of up to $100,000 per day for violations. Individuals who knowingly issue stablecoins in the United States without being a permitted payment stablecoin issuer face up to five years’ imprisonment and fines up to $1 million for each violation. Additionally, individuals with certain felony convictions are prohibited from serving as officers or directors of a PPSI, and violations of that prohibition can result in imprisonment for up to five years. The Act expressly gives regulators discretion to refer violations of the Act to the Attorney General.

Modernizing anti-money laundering and financial crimes compliance

The GENIUS Act places a strong emphasis on leveraging blockchain technology and AI to modernize the detection of illicit financial activity involving digital assets. The Act mandates that the Secretary of the Treasury initiate a public comment period to gather insights on how regulated financial institutions are using or could use innovative tools—particularly blockchain and AI—to detect money laundering and related crimes. Blockchain technology is highlighted for its potential in transaction monitoring and transparency, especially in tracking digital asset flows and identifying suspicious patterns.

Rulemaking timeline

The Act mandates that all primary federal payment stablecoin regulators, the Secretary of the Treasury, and state payment stablecoin regulators must promulgate regulations to implement the Act within one year of its enactment (July 18, 2026). These regulations must be issued through a notice-and-comment process. Additionally, within 180 days of the Act’s effective date, the OCC, Federal Deposit Insurance Corporation, and Board of Governors of the Federal Reserve System shall submit a report to the Senate Committee on Banking, Housing, and Urban Affairs and the House Committee on Financial Services that confirms and describes the regulations necessary to carry out this Act.

Other Impending Crypto Legislation

The GENIUS Act is momentous for stablecoin issuers, but it does not resolve a number of crypto-native issues, which are the subject of a broader market structure bill known as the Digital Asset Market Clarity Act of 2025 (CLARITY Act). The CLARITY Act passed the House with broad bipartisan support, and a version is currently under Senate consideration. While the GENIUS Act focused narrowly on regulating stablecoin issuers, the CLARITY Act seeks to establish a robust regulatory framework for all digital assets and define the roles of the Securities and Exchange Commission and Commodity Futures Trading Commission in policing the digital asset markets. Most notably, for the first time, the CLARITY Act attempts to classify digital assets based on their characteristics, such as decentralization and blockchain maturity, with a goal of reducing regulatory uncertainty and fostering innovation in the cryptocurrency industry. Senator Tim Scott (R-SC), chair of the Senate Banking Committee, has made several public statements on the timeline for consideration of the CLARITY Act, with committee markup expected in September and full Senate action possible by late fall.

Conclusion

The GENIUS Act establishes a robust framework for the issuance and oversight of payment stablecoins in the United States. It sets clear standards to ensure transparency for the backing of permitted payment stablecoins, and it requires issuers, like traditional financial institutions, to quickly establish robust compliance programs to combat illicit uses of their stablecoins. With its strong bipartisan backing and goals of financial stability, consumer protection, and global competitiveness, the Act could lay the groundwork for a more transparent and trustworthy digital asset ecosystem.

Ransomware Fusion Center

Stay ahead of evolving ransomware threats with Alston & Bird’s Ransomware Fusion Center. Our Privacy, Cyber & Data Strategy Team offers comprehensive resources and expert guidance to help your organization prepare for and respond to ransomware incidents. Visit Alston & Bird’s Ransomware Fusion Center to learn more and access our tools.

Originally published July 24, 2025.

If you have any questions, or would like additional information, please contact one of the attorneys on our Privacy, Cyber & Data Strategy team.

You can subscribe to future advisories and other Alston & Bird publications by completing our publications subscription form.