.

Client Alert

Launching Agentic AI in an Uncertain U.S. Regulatory Landscape

January 28, 2025

by Sarah Robertson

Download as a PDF

Share this page

Are you ready to begin adding AI Agents to your human teams? You will soon be getting requests to do so. While business teams will be wowed by what AI Agents and their subagents can do, the artificial intelligence regulatory environment is increasingly uncertain, and we advise caution.

Within its first few days, the new administration revoked[1] the 2023 Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence[2] (the “Old AI Executive Order”) – which had been intended to mitigate risks associated with AI technologies by imposing safety guidelines such as a watermarking system for AI-generated content. Next, the president issued a new executive order Removing Barriers to American Leadership in Artificial Intelligence[3] (the “New AI Executive Order”), providing White House staff 180 days to develop a plan “to sustain and enhance America’s global AI dominance in order to promote human flourishing, economic competitiveness, and national security.”[4] Under the New AI Executive Order, federal agencies are required to immediately review and suspend, revise, or rescind any prior governance and safety measures undertaken to comply with the Old AI Executive Order that are not consistent with the New AI Executive Order.

Meanwhile, on January 21, the president announced a new $500 billion private sector deal, “The Stargate Project,”[5] which is a new joint venture that relies on SoftBank, OpenAI, Oracle, and MGX as initial equity funders in a bid to rapidly expand U.S. AI infrastructure by building massive new data centers, first in Texas then in other potential sites across the country. In the same week, leading big tech companies including Oracle and OpenAI each announced[6] the next phase of the AI revolution with the release of new AI Agents designed to autonomously handle specific tasks and ultimately “join the workforce and materially change the output of companies.”[7]

What Are AI Agents?

While traditional AI chatbots using large language models (“LLMs”) are designed to respond to users within the AI system, the latest generation of advanced AI Agents (aka AI Super-Agents or Agentic AI) are now empowered to interact with computer, network, and internet environments to automate tasks. Open AI states that its new AI Agent, Operator, “can be asked to handle a wide variety of repetitive browser tasks such as filling out forms, ordering groceries, and even creating memes.”[8] Accordingly, Operator “can ‘see’ (through screenshots) and ‘interact’ (using all the actions a mouse and keyboard allow) with a browser, enabling it to take action on the web without requiring custom API integrations.”[9] By combining the power of LLMs with new technologies such as the Oracle Cloud Infrastructure and retrieval-augmented generation, Oracle’s AI Agents can interact with enterprise data and can apparently be used by companies to recruit qualified job candidates, perform complex customer data analytics, optimize call centers, and expedite legal, financial, and academic research tasks.

AI Agents can be trained or instructed to follow the values of a user or business. Whether that can be consistently implemented remains to be seen. In testing, some AI systems have been shown to engage in “scheming” where they change results or take unexpected actions toward an ultimate goal set by the user. Although the usefulness of these technologies is undeniable, companies looking to deploy AI Agents should be aware of the legal risks and pitfalls that may accompany the use of these revolutionary new tools.

Legal Challenges and Mitigation Strategies

  • AI Agency Liability. In July 2024, a California district court allowed a case against HR and finance platform Workday to proceed, stating that an employer’s use of Workday’s AI-powered HR-screening algorithm may create direct liability for both the employer and Workday under the theory of agency liability.[10] In this hiring discrimination case, the plaintiff alleged that his prospective employer delegated “traditional hiring functions, including rejecting applicants, to the algorithmic decision-making tools provided by Workday.”[11] The court found that by designing an AI technology to make decisions that would normally be made by a live employee, Workday should be treated as an agent of the employer for purposes of liability so long as the employer actually relied on the AI technology in its hiring process. While this case is still making its way through the courts, AI Agent vendors and deployers could subsequently be exposed to both civil and criminal liability based on the actions of the AI Agent, a theory that had previously only been applied by the courts to the actions of live humans.
  • AI Product Liability. Given the increased responsibility and autonomy being granted to AI Agents, it is possible that developers and deployers of AI Agents could also be required to contend with product liability claims. Manufacturers and sellers can be held responsible for injuries caused by defective or unreasonably dangerous products. If an AI Agent makes a poor and costly decision, a plaintiff may claim that the AI Agent was defectively designed and/or that the developer failed to adequately warn the plaintiff of the AI Agent’s limitations. This theory of liability is currently being tested in cases against AI-chatbot platform Character.AI. In the fall of 2024, the mother of a deceased minor in Florida brought a product liability case against the developers of Character.AI – an AI-chatbot role-playing platform that allows users to create and converse with AI-powered characters – alleging that her son died by suicide after becoming harmfully dependent on his relationship with his Character.AI companion “Dany.”[12] In December, parents in Texas brought additional product liability claims, stating that, without warning, Character.AI chatbots exposed their children to hypersexualized content, self-harm, and violent behaviors, with one chatbot allegedly encouraging a child to kill his parents when they tried to limit his screen time.[13]
  • AI Contractual Considerations. Businesses seeking to onboard AI Agents should implement clear contractual provisions to allocate and manage risk. When reviewing warranties, limitations of liability, and indemnification clauses, companies should want to know if and to what extent the AI Agent vendor will indemnify the company for an AI Agent’s decision making, especially if those decisions are illegal or cause harm to end-users or others. Businesses should also consider who is responsible for training employees to use AI Agents safely and whether that training relates to the AI provider’s contractual liability. In addition to requiring compliance with emerging state and international AI regulations such as the Colorado AI Act[14] and the EU AI Act,[15] companies should also carefully contemplate intellectual property (“IP”) and data ownership, rules for algorithmic training inputs, and customization of standard liability shifting terms. Whereas typical software-as-a-service contracts usually address ownership of software and/or underlying data, AI Agent vendor contracts should also address IP-ownership of AI-generated content such as images, text, and even new software. Regarding algorithmic training inputs, companies must decide whether the AI Agent’s algorithm can train on company data, and if so, how such company data must be protected and stored. In addition, businesses should be wary of AI Agents violating third party terms of service which may prohibit access and use by bots. In non-negotiated use of Agentic AI, such as when a company deploys an off-the-shelf AI Agent from one of the major providers, companies should similarly review the provisions outlined above and weigh the ultimate risk of deploying this novel technology.
  • AI Privacy and Cybersecurity Considerations. Using an AI Agent may result in the processing of enormous amounts of personal information as defined by state, federal, and international data privacy laws. For example, AI Agents designed to read and prepare automated responses to emails may digest any personal information contained in a user’s inbox, while an AI Agent designed to make investment decisions will process sensitive financial information about the user. By collecting such large amounts of personal information, companies may grow their cybersecurity attack target, as bad actors are typically attracted to companies known to collect large amounts of detailed and/or sensitive personal information. Furthermore, by deploying an AI Agent, a company in scope for comprehensive state privacy laws, such as the California Consumer Privacy Act, may be required to offer consumers the opportunity to opt-out of the AI-Agent’s automated decision making, while also disclosing to its consumers how their personal information is collected and used by the AI Agent, including whether their personal information subsequently trains the AI Agent’s algorithm. AI Agents that are not adequately directed and supervised could also perpetrate scams, develop vulnerable software code, or cause cybersecurity incidents, so human supervision and real-time monitoring will be essential to reduce legal risk, especially with initial uses of AI Agents.

With the launch of AI Agents, concerns about AI scheming have become immediate. Adding AI Agents to your teams may have serious, unintended consequences and should involve significant testing and implementation of controls prior to and during use. Businesses will soon be training AI Agents in their corporate values and adding AI Agents to their teams. The way we work will never be the same.

Dorsey’s Privacy, Cybersecurity & Social Media, Technology Commerce, and Intellectual Property groups are monitoring technological developments and regulatory updates and are here to help.

 


[4] Id.

[11] Id.

Firm Highlights

Insights

Minnesota Consumer-Fraud Enforcement: Disclaimers Won’t Cure Deception, and Penalties Can Add Up

Misleading advertising claims can lead to costly consequences. In State of Minnesota ex rel. Ellison v. Wall & Associates, Inc., No. A25-2064, 2026 WL 3118291 (Minn. App. Oct. 19, 2026), the Minnesota Court of Appeals recently affirmed a bench trial verdict against a tax-relief company and two executives for violations of the Minnesota Consumer Fraud Act (“MCFA”) and the Minnesota Deceptive Trade Practices Act (“MDTPA”). The decision, which affirmed a substantial restitution and civil penalty judgment, underscores the broad power of the Minnesota Attorney General to act against deceptive sales practices. The decision also offers practical lessons for businesses seeking to reduce liability risks in consumer-facing communications. Background Wall & Associates (“W&A”), a Virginia-based tax-relief company, marketed assistance to taxpayers seeking relief from outstanding federal and state tax liabilities, often with offers in compromise—agreements to settle tax liabilities for less than the amount owed. In its extensive advertising campaigns, W&A represented that its average client received a 90% reduction in their tax debt, that it was a local company with Minnesota offices, that it provided legal services, and that it would “Solve ALL Tax Problems.” In reality, W&A’s promises bore little resemblance to what many clients received. Most of W&A’s clients saw just a 9.5% reduction in federal tax debt and 1.2% in state tax debt. W&A had no permanent Minnesota offices. And W&A provided neither legal representation nor assistance with filing tax returns—those services required referrals to separate firms for an additional fee. All the while, W&A clients paid upfront and recurring fees, including during periods when W&A did no work on their cases. Minnesota Attorney General Keith Ellison sued W&A, its founder and owner Ernest Kenneth Wall, and its president and CEO Patrick Mark Yates, alleging violations of, among other laws, the MCFA and the MDTPA. Following a 12-day bench trial, the district court determined that W&A violated the MCFA and MDTPA through its false and embellished advertising. The district court also held Wall and Yates individually liable for those violations because they directed, participated in, and acquiesced in the violations. The district court ordered injunctive relief, $2.7 million in restitution (with Wall and Yates jointly and severally liable for $1.3 million), and civil penalties of $786,517 against W&A, $415,500 against Wall, and $207,750 against Yates. The district court also awarded over $2.5 million in attorneys’ fees and costs to the state. W&A, Wall, and Yates appealed to the Minnesota Court of Appeals, which affirmed the district court’s decision on October 19, 2026. Although W&A, Wall, and Yates challenged the judgment on numerous grounds, several highlights of the decision provide important takeaways for consumer-facing businesses. Individual Reliance Is Not Required for Statutory Liability W&A argued that Attorney General Ellison had to prove a causal nexus between each violation of the MCFA and MDTPA and the injury to a particular consumer to establish liability. In W&A’s view, this required proof that each injured consumer had personally relied on W&A’s false representation. The Court of Appeals rejected this argument based on the MCFA’s and MDTPA’s plain language, which requires no proof of actual deception or confusion by consumers. The Court of Appeals finally pointed to the Minnesota Supreme Court’s decision in State v. Minnesota School of Business, Inc., which held that when a defendant’s misrepresentations are directed at and affected “a broad group of consumers, proof of direct individual reliance is not required to establish a causal nexus between MCFA violations and the harm suffered by consumers.” 935 N.W.2d 124, 135 (Minn. 2019). The Court of Appeals affirmed, however, that a causal nexus is necessary to support an award of damages under the MCFA and MDTPA. In making that determination, district courts look to various factors, including whether the misrepresentations were “longstanding, pervasive, and widespread in communications directed to consumers,” whether the seller intended and understood that consumers would rely on the misrepresentations, and whether the information was of a kind on which consumers would typically rely. The Court of Appeals’ decision continues Minnesota School of Business’s broad interpretation of liability under the MCFA and the MDTPA. Coordinated marketing campaigns to a “broad group of consumers” may expose businesses to MCFA and MDTPA liability without customer-by-customer proof of injury. That said, a causal nexus remains an element of an MCFA and MDTPA claim, and businesses involved in litigation should build a clear record at trial and on appeal showing how each of the relevant factors bears on the causal-nexus inquiry. Contractual Disclaimers Are Not an Automatic Defense to a Misleading Pitch W&A next argued that disclaimers in its client contracts prevented consumers from claiming they had been misled by its advertising. Reiterating that individual reliance is not required under the MCFA and MDTPA, the Court of Appeals held that a contractual disclaimer does not necessarily defeat a claim based on a conflicting oral representation. And on the facts of this case, the Court of Appeals agreed with the district court that W&A’s contractual disclaimers were insufficient given the extensive misrepresentations by W&A and the unequal bargaining power between it and its customers. This holding confirms that contractual disclaimers are not a cure-all for misleading advertising or sales practices. Businesses should review the entire customer journey, including pre-contract representations, as those representations may survive a contractual disclaimer to the contrary. And if faced with litigation, businesses should build a strong record, both at trial and on appeal, showing why any disclaimer is effective in context. The State Has Broad Authority to Seek Restitution W&A also challenged Attorney General Ellison’s authority to seek restitution under the state’s parens patriae authority to protect residents’ collective interests, arguing that injury to only 228 Minnesotans did not affect a sufficiently substantial segment of the population. The Court of Appeals rejected a simple headcount approach, instead explaining that the state had a legitimate interest in protecting residents’ economic well-being as a whole. In any event, the Court of Appeals held that if restitution was not appropriate under the state’s parens patriae authority, then it would still be appropriate under the Attorney General’s enforcement powers under Minn. Stat. § 8.31. The decision reinforces that the broader and more sustained the deceptive advertising, the stronger the case for restitution to protect Minnesotans’ economic interests. Civil Penalties Accrue Per Violation, Not Per Defendant The Court of Appeals finally rejected W&A’s claim that the Attorney General’s enforcement powers under Minn. Stat. § 8.31 permit only civil penalties of up to $25,000 for each defendant. Rather, the Court of Appeals explained that penalties are limited to $25,000 per violation. Multiple violations can therefore produce substantially greater penalties. The Court of Appeals also rejected the argument that restitution made penalties unnecessary: repaying consumers does not eliminate the need to address gains from deceptive conduct and deter future violations. Final Takeaways Align every consumer-facing communication. Broad and extensive marketing campaigns can expose businesses to stronger enforcement standards under the MCFA and MDTPA, making it critical that consumer communications remain accurate and consistent from the first advertisement through the final sale. Plan for layered enforcement exposure. Penalties under the MCFA and MDTPA can add up quickly: W&A, Wall, and Yates ultimately faced more than $6 million in civil penalties, restitution, and attorneys’ fees. The decision continues a line of Minnesota cases recognizing the Attorney General’s broad consumer-protection authority. Details matter. Although the Attorney General’s enforcement power is broad, it is not boundless. As the Court of Appeals explained, issues of causation, the effectiveness of contractual disclaimers, and the availability of restitution and civil penalties all turn on the specific facts of the case. Effective advocacy at trial and on appeal is critical to developing the facts relevant to liability and damages.

Insights

Observations on the SEC’s 2027 Examination Priorities

The SEC’s Division of Examinations (the “Division”) recently announced its examination priorities (“Exam Priorities”) for fiscal year 2027.1 Although many of the Division’s core areas of focus remain consistent with fiscal year 2026, the 2027 Exam Priorities include notable changes and refinements, which we discuss below, along with other key 2027 examination priorities. Dorsey Observations Renewed focus on economic conflicts of interest. After certain economic-incentive topics received less emphasis in the 2026 Exam Priorities2, the Division has again expressly identified economic conflicts as an examination focus. The 2027 Exam Priorities indicate that the Division will prioritize reviewing advisers’ fiduciary obligations relating to economic conflicts of interest raised by fee rebates, fee and expense calculations, cash management, quid pro quo arrangements, and deposits and withdrawals. Greater emphasis on private-fund practices. Building on the private-fund items in the 2026 Exam Priorities, the 2027 Exam Priorities indicate that the Division will assess private-fund disclosures and controls relating to investment allocations, leverage, affiliated service providers, and fees and expenses. The Division also identified investments in private companies through special purpose vehicles (“SPVs”) and loans provided by private funds and business development companies (“BDCs”) as products the Division will scrutinize. Advisers, particularly private-fund advisers, should evaluate whether their disclosures and compliance programs address these areas. Expanded focus on AI-related cybersecurity risks. While the 2026 Exam Priorities focused on advisers’ use of and representations regarding artificial intelligence (“AI”), the 2027 Exam Priorities specifically highlight AI-enabled cybersecurity threats, including AI-generated phishing and synthetic or “deepfake” voice and video used for social engineering. The shift signals that registrants should consider whether their cybersecurity policies and controls adequately address these evolving threats. New focus on proxy advisory firms. The 2027 Exam Priorities state that the Division will examine advisers that rely on proxy advisory firms, including how those advisers evaluate the firms’ recommendations, assess the bases for those recommendations, identify potential conflicts of interest, and determine whether that reliance satisfies applicable legal obligations. Advisers that rely on proxy advisory firms may wish to consider whether their compliance programs address these areas. Updated focus on Regulation S-P and Regulation S-ID. The 2026 Exam Priorities stated that the Division would engage firms during examinations regarding their progress in preparing incident response programs ahead of the compliance dates for the amendments to Regulation S-P and would examine compliance with the amended rule after those dates. The 2027 Exam Priorities indicate that examinations will focus on whether firms have developed, implemented, and maintained reasonably designed policies and procedures addressing administrative, technical, and physical safeguards to protect customer information and that, in accordance with the amended rule, staff will evaluate response program policies and procedures for unauthorized access to or use of customer information, including customer notification procedures and oversight of service providers. With respect to Regulation S-ID, the 2027 Exam Priorities indicate that examiners will assess whether firms are updating their Identity Theft Prevention Program to address emerging risks, such as those related to AI and stolen credentials. Advisers may wish to consider whether their compliance programs address these areas, as applicable. Summary of Other 2027 Examination Priorities Beyond the developments discussed above, many of the Division’s core examination priorities remain consistent with fiscal year 2026. These include investment adviser fiduciary obligations and compliance programs, examinations of never-examined and recently registered advisers, and cybersecurity and operational resiliency. The Exam Priorities are not exhaustive, and examinations will vary based on each registrant’s business, activities and risks. Advisers should consider the 2027 Exam Priorities in reviewing their compliance programs and preparing for potential examinations, including by: Reviewing risk disclosures and controls. The Division states that, when examining advisers to private funds, it will assess whether investment-related risk disclosures are consistent with fiduciary standards and whether compliance policies, procedures, and controls are designed to address both fund-level and investment-level risks. Reviewing disclosures on allocations, leverage, and affiliated service providers. The 2027 Exam Priorities identify disclosures related to certain investments and allocations, the use of debt and fund-level lines of credit, and the use of affiliated service providers as examples. Confirming the accuracy of fee and expense calculations. The Division states that it will assess the accuracy of fee and expense calculations at both the fund and investment levels, and it identifies fee rebates, cash management, quid pro quo arrangements, deposits, and withdrawals as areas in which economic conflicts of interest may affect the impartiality of advice. Reviewing recommendations of illiquid or hard-to-trade products. The Division states that it will prioritize advisers that recommend products that are difficult to trade, have unusual features, or are illiquid, including loans provided by private funds or BDCs and investments in private companies through SPVs. Confirming that the compliance program is implemented and reviewed. The Division states that it will review compliance program components such as marketing, valuation, and custody, as appropriate, and will typically analyze the annual compliance review. It will also focus on compliance practices when advisers have started advising new asset classes or clients or grown through mergers or acquisitions. 1 SEC Division of Examinations 2027 Examination Priorities. 2 See Dorsey’s Client Alert on the SEC’s fiscal year 2026 exam priorities.