Key insights:

• • • AI is supercharging old fraud schemesĚý— By making synthetic identities, deepfake scams, and customer fraud faster, more credible, and harder to detect, AI is amplifying fraud and crime.

    • The real vulnerability may be internal silosĚý— Institutions need to be on the lookout, because what looks like a credit loss, an HR issue, or a payment request may actually be part of a wider multi-vector AI-enabled attack.

    • Institutions already have the tools to respondĚý— Through KYC and internal and behavioral data, financial institutions have the ability to respond to fraud threats — but only if teams connect and act together.

Fraud and crime existed long before AI, of course, but today’s technology delivers an acceleration in speed, scale, and success rate for fraudsters, resulting in billions of dollars in losses for victims. AI-enabled frauds on financial institutions by 2027 in the United States alone, and of detected fraud attempts on financial institutions use AI – and of these, 29% are successful.

To respond effectively to these threats, institutions need to implement a unified response that brings together departments that may not traditionally be partners. This cross-functional coordination should include not only the institution’s fraud and financial crime risk teams but also its credit risk, cybersecurity, and human resources functions.

And this response is critical, because today, financial institutions are being targeted by multiple types of AI-enabled attacks, including tactics such as:

• • • use of synthetic identities to circumvent know your customer/customer due diligence (KYC/CDD) controls and perpetrate fraud or launder money;
    • use of deepfake identities to gain employment, particularly by North Korean IT workers;
    • AI-enhanced “CEO frauds” to deceive staff into taking unauthorized actions; and
    • Bank customers may be targeted by fraud too, presenting further risk to financial institutions.

Let’s look at these threat vectors individually:

Vector 1: Synthetic identities and KYC/CDD

Synthetic identities can be entirely fabricated or may use combinations of real and fabricated personal information to create a new identity. For example, a fraudster may construct a synthetic identity using a Social Security number exposed during a data breach combined with an AI-generated passport.

This threat is real and happening now: identifies that criminals have already used AI to successfully open accounts using falsified documents, photographs, and videos. And according to , synthetic identities were used to open as many as 3% of US bank accounts, representing millions of identities. Not surprisingly, these illicit accounts are used to commit fraud and launder the proceeds of money laundering.

Vector 2: North Korean IT workers

North Korean individuals have successfully gained employment as remote IT workers at American companies, often passing themselves off as US nationals using AI-generated face-swapping technology combined with proxy computers and false identity documents. North Korean IT workers are almost $800 million annually for the regime.

Institutions deceived into employing these workers are not only against North Korea, but they are also exposing commercially sensitive data and systems to an adversary state, increasing the possibility of theft, cyber-attacks, and extortion.

Vector 3: CEO Fraud

A “CEO fraud” is a cybercrime in which an attacker impersonates an executive to deceive an employee into taking actions such as sending unauthorized wire transfers or disclosing sensitive information. AI accelerates these frauds by making them more personalized and credible.

In one of the more well-known examples, in an AI-enhanced CEO fraud in 2024 after the fraudster impersonated Arup Engineering’s CFO and requested a staff member to make several financial transfers. The criminals added credibility to the fraud by using a in which the target recognized many of their colleagues – unfortunately, all of them were deepfakes.

Vector 4: Frauds targeting customers

Where customers are targets, AI provides the scale, speed, and personalization to allow illicit actors to deliver individualized fraud. For example, whereas romance scams previously used repetitive scripts and re-used the same images of the romantic “partner,” fraudsters can now use AI-generated messages, images, or videos, continuously adapting the execution of the scam to the target’s responses and behaviors.

Creating a cross-functional and unified response

The examples above demonstrate the diverse and highly sophisticated uses of AI by illicit actors, both adversary states and criminal networks. Detecting and responding to these illicit activities requires joint action between teams that may not traditionally work closely together.

For example, if an account holder fails to repay a loan, the credit team may consider it to be a default by a legitimate customer and write it off as a credit loss. However, if the account was opened using a synthetic identity, investigation may reveal other accounts that share similar customer data points or transactional patterns. This could reveal a network of accounts that are perpetrating a fraud or money-laundering scheme. To detect and respond effectively, joint action is needed between KYC/CDD on-boarding teams, financial crime investigators, and fraud and credit risk professionals.

Alternatively, for HR teams to effectively identify use of face-swapping videos during a hiring process, knowledge from the organization’s cybersecurity team, especially of deepfake indicators, would be valuable. If a North Korea IT worker is hired and only later identified, cybersecurity and sanctions teams must be involved in the response to mitigate data, network, and compliance exposures.

Detecting and responding to all illicit activities requires joint action between teams that may not traditionally work closely together.

Finally, all staff may be targeted by deepfake fraud, but those in senior positions or departments with financial authority are the most vulnerable. This means it is essential for institutions to deliver employee training using real-life case studies, “near misses,” and scenarios drawn from across the institution and industry. This type of training will increase vigilance and minimize the likelihood of a successful attack.

For customers, financial institutions are well-positioned to identify indicators of fraud due to their extensive datasets of KYC/CDD records, transactional, and behavioral information. Institutions should enhance their customer relationships (as well as meet applicable regulatory requirements) by taking proactive measures to inform and protect their customers.

While AI has accelerated fraud and crime, financial institutions also hold valuable and relevant assets: the knowledge distributed across their cybersecurity, HR, credit risk, financial crime compliance, fraud, and KYC/CDD teams. By connecting these teams together, even in contexts in which these departments have not traditionally been partners, institutions will be well-positioned to protect both themselves and their customers from illicit actors’ sophisticated AI-enabled threats.

You can learn more about the fraud-fighting challenges faced by financial institutions and other organizations here

Key highlights:

• • • The theory-practice gap is now an AI-era crisis— Integrating legal training with hands-on pro bono experience is the future of legal education.

    • A collaborative model merges learning and doing into a single platform— The model connects law students with vetted pro bono opportunities from legal services organizations, while also offering targeted, skills-based training at the moment students step into those matters.

    • Pro bono work is uniquely suited for responsible AI training— On-demand programs led by expert faculty are available to help students sharpen pro bono skills, understand the use of AI in today’s legal practice, and stay on top of developments in numerous industry and practice areas.

Legal education has operated on a familiar, decades-long divide that saw students spend their first years learning the law in the classroom and then after graduation, gaining substantive experience practicing the law in the real world. This gap has always been costly for both students and legal employers, and now it’s emerging as untenable in an era in which AI is rapidly reshaping what junior lawyers do.

Pro bono and skills training close this gap

A new partnership between , a pro bono management platform, and the (PLI), a nonprofit provider of learning resources for legal professionals, is designed to close this gap while showing something larger about where legal education must go.

The partnership is designed to equip students with on-demand, actionable training that supports effective pro bono engagement by offering access to PLI’s training programs directly through Paladin’s platform. Since launching with 30 law schools in August 2025, students have signed up for thousands of pro bono cases through the platform, according to , Co-founder and CEO of Paladin.

For years, experiential learning in law schools was something students had to piece together on their own by hunting across spreadsheets, clinic listings, and externship postings for opportunities, says Sonday, adding that too often students were given little guidance on what they were walking into.

The partnership is designed to equip students with on-demand, actionable training that supports effective pro bono engagement

“What’s fundamentally different is the integration and centralization of learning and doing,” Sonday explains. “Historically, legal education has separated theory, training, and practice.” Now, she notes, a student can learn a concept, build confidence through targeted training, and apply it in a real-world setting within a short amount of time.

, Chief Strategy Officer at PLI, describes the experience from the student’s perspective: “When a first-year logs into the Paladin platform, they are not thrown into the deep end. Instead, they can access skills-based programs, such as a PLI program specifically on how to interview a pro bono client before they ever sit across from someone in need. This leads to a better experience for the student, the law school, and especially for the client.”

Pro bono work suited to responsible AI training

The urgency behind this partnership is inseparable from the impact AI is having on the entry-level legal market.

“We’re already seeing AI reduce the time spent on tasks like initial legal research, document review, drafting memos, and summarizing case law,” Sonday says. “This is work that has traditionally formed the foundation of junior associate training.” The skills AI cannot replicate — such as judgment, issue spotting in ambiguous situations, client communication, and ethical decision-making — are what students need to develop deliberately earlier in their legal careers.

Indeed, those human skills are essential to the effective use of AI, Talmage says. The lawyer of the future will be a strategic advisor and creative problem solver, which are the very attorney roles that AI cannot fill, she explains, adding that those must be cultivated through experience. “You always need to be questioning and verifying and authenticating — and that’s generally a lawyer’s role.”

For years, experiential learning in law schools was something students had to piece together on their own by hunting across spreadsheets, clinic listings, and externship postings for opportunities.

There is a particular logic as to why pro bono work is the right fit for learning to use AI responsibly. Pro bono is “a built-in, humans-in-the-loop model” in which students are always supervised by attorneys, Sonday says. And this supervision creates a structured environment in which to learn how to use AI tools, apply them to real matters, get feedback, and iterate. The result, Sonday argues, will be more attorneys who are AI-fluent early on and throughout their careers.

A message to law school leaders

For law school leaders, both Sonday and Talmage highlight that AI use has already changed the legal profession. The choice then for law schools is whether they evolve by design or by default.

Students know the legal profession has changed and so do employers, CLE providers, and clients, Talmage explains.

Sonday agrees. “The pace of change in the legal profession is accelerating, and students need to be prepared not just for the law today, but also for the practice of law in the future,” she says. “Integrating pro bono platforms and AI-specific training aligns legal education with reality.”

The Paladin/PLI partnership offers a blueprint for what legal education must become in the future, transforming itself into a space that’s grounded in applied legal knowledge, human-supervised, and AI-informed. Indeed, the best way to train the next generation of lawyers is to give them real clients, real cases, and real responsibility while they still have room to grow.

You can find more about the challenges facing law schools and legal education here

Key insights:

• • • AI is changing legal’s role, not just its workload — Going forward, AI will do more than automate routine tasks, it also will help in-house legal teams become more strategic business partners.

    • The 5 archetypes make the transformation concrete — There are five practical ways in which AI could reshape legal work, including automation, stronger advising, better collaboration, and global scale.

    • Every organization’s AI transformation will be different — ¶¶ŇőłÉÄę’ own legal transformation journey shows the common and unique aspects of this process.

Beyond the automation, productivity boosts, or the now-familiar promise of doing more with less, the question over how AI will really transform the work that corporate legal departments do on a daily basis, has yet to be truly answered.

To deepen our understanding of where in-house legal is really heading next, Norie Campbell, ¶¶ŇőłÉÄę Chief Legal Officer, and Lizzy Duffy, a Senior Director of the Thomson Reuters Institute, produced a new feature article, The 2030 legal department: 5 ways AI will transform how in-house teams workĚýthat steps back from the day-to-day noise around AI and asks the bigger, more interesting question: “What is the legal function actually becoming?”

Importantly, the article recognizes that in-house legal teams are navigating real constraints around time, budget, and clarity even as expectations continue to evolve. It also acknowledges how GCs are balancing rising demands with a growing focus on efficiency, while also working to define what effective and meaningful AI adoption should look like for their teams.

Indeed, this human pressure is one of the most compelling aspects to the questions corporate law departments are facing today, and it reverberates beyond a simple theory of AI in legal to really reflect a profession at a turning point.

The five archetypes

The feature also lays out five archetypes — distinct models for how AI could reshape legal work, from high-volume automation to better strategic advising, stronger business partnering, smarter collaboration with outside counsel, and truly global leverage across teams and languages.

By referencing these five archetypes, legal department leaders can start asking where their own teams fit, and what they need to do to get better prepared for the AI-driven legal future of 2030.

These archetypes cover everything from deciding on the best ways to leverage AI-led automation to helping legal teams become more proactive strategic advisers. The archetypes also detail how to foster collaboration that can allow other corporate functions to act more confidently without constant legal intervention. And how to use AI to reduce barriers caused by language and time zones, enabling multinational legal teams to work more effectively across geographies.

By referencing these five archetypes, legal department leaders can start asking where their own teams fit, and what they need to do to get better prepared for the AI-driven legal future of 2030.

¶¶ŇőłÉÄę’ own journey

This feature article also builds a practical, grounded picture of the future from inside ¶¶ŇőłÉÄę’ own General Counsel’s Office (GCO), showing readers a transformation that’s already taking shape.

This insider perspective offers a front-row look at how one GCO is trying to move from experimentation to real transformation and tells a bigger story than technology alone. Today’s transformation of the corporate legal department is really about leadership, ambition, and the choices department leaders need to make now if they want to stay relevant by 2030.

More than anything, the feature article stresses that adopting AI tools is not the same as true transformation. To move beyond incremental gains, legal departments must redesign workflows, improve data infrastructure, invest in training, and hire for adaptability and technical literacy. Ultimately, the central message is that efficiency is only a by-product — the real challenge is deciding what kind of legal function an organization will need in 2030 and how to start building toward that vision now.

You can access the full feature article, The 2030 legal department: 5 ways AI will transform how in-house teams work here

Key highlights:

• • • CurriculumĚýredesign must start now — One law school’s approach illustrates the necessity of mapping the entire curriculum to identify which skills to preserve, evolve, or build from scratch.

    • Training faculty in AI use is critical — Faculty AI training should be a multi-layered approach including hands-on training with specialized legal AI tools, guidance on redesigning curricula, and more.

    • AI simulations may be the key — Law school leaders need to act now by experimenting with small pilot projects and building simulation-based learning tools to replace the developmental depth that once came naturally in the first years of practice.

The debate about AI consuming most of the work that teaches essential lawyering skills to junior attorneys is forcing a reckoning with the long-held assumption that law schools were never designed to produce practice-ready lawyers and that it was always the profession’s job.

Indeed, AI is forcing that uncomfortable truth into the open faster than anyone anticipated because essential lawyering work — the document review, contract markup, research memo creation — dictated how a junior lawyer learned to spot the issue buried on page 47, to sense when a clause was off, and to develop the instinct that no classroom can fully replicate. Now, as more law firms deploy AI to handle precisely those entry-level tasks, the organic training moments that used to define the first two to three years of legal practice are evaporating.

, Executive Dean, Faculty of Law at Bond University, and Co-Chair of the Council of Australian Law Deans, says he sees where this is leading. The ultimate results will be firms hiring fewer junior lawyers today because AI has taken over that entry-level work, James explains, adding that means there will simply be no pipeline of mid-level, experienced lawyers to draw from in three to five years. Indeed, this is a slow-moving crisis, already in motion, and yet to fully arrive.

This crisis lands at the center of what the AI and Future of Legal Practice (AIFLP) initiative exists to address because at the core of this crisis is what does being job-ready really means when the job itself is being redefined. Answering this question requires law schools, law firms, licensing bodies, and technologists to do something they have historically struggled to do — that is to think and act collaboratively.

Rethinking the curriculum before AI does it for you

leads IE Law School’s AI initiative and is steering the school’s efforts to embed AI across the curriculum. To do so effectively, her approach requires going back to a broader set of foundational questions in legal education such as: For what is legal education meant to prepare students? How do students learn to develop legal judgment? What makes legal advice genuinely valuable? And what skills are essential to deliver that value in an AI-enabled profession?

“Layering AI tools on top of an unchanged curriculum serves no one,” Perez-Llorca explains, adding that without answers to the fundamental questions, “you are just adding technology to a structure that was never designed to handle it.”

Check out how one law school professor is building AI simulation tools

IE law school is currently mapping its entire curriculum to determine which skills need to be preserved, which need to evolve, and which need to be built from scratch, while also using the AI-boosted curriculum to train faculty. Perez-Llorca describes the school’s faculty AI training as a multi-layered approach encompassing university-wide LLM training, substantive AI law curriculum review, hands-on training with specialized legal AI tools, guidance on redesigning curricula, and assessments to reflect students’ growing AI proficiency. Before students can be taught with AI, professors need to understand the tools themselves and how to use them in teaching, in simulation, and in assessment, she adds.

An AI tutor that meets students where they are

Bond University’s James says he has spent the last several months building an AI tutor designed to walk students through course material the way a patient, attentive instructor would. His vision for the AI teaching assistant supports the professor meeting students where they are. “It [the AI tutor] introduces the week’s topic, outlines learning outcomes, guides students through the readings, checks comprehension with short quizzes, and then adapts in real time based on how the student responds,” James explains, adding that the AI tutor will pull any student who is struggling deeper into the material until the learning outcome is achieved. “The conversation never stops until the learning does.”

However, James is careful to draw a clear distinction about what the tutor replaces and what it does not, stressing that AI is a substitute for the lecture recording, the static reading list, or the passive video watched at midnight before an exam — but it chiefly exists to support the law professor. This approach frees up class time, turning it from content delivery to more meaningful the time between the human instructor and students, he adds.

Act by design or default

The approaches by both Perez-Llorca and James point to a way to address the question of disappearing tasks that teach essential lawyering skills as well as shift the center of gravity in legal education toward ways to foster developmental skills and legal judgment. Indeed, inertia is not a strategy, and law school deans and associate deans can be at the forefront of this fight by taking decisive action, including:

• • • Experiment freely — Investigate with AI on your own by starting small with a pilot project.
    • Strategically assign where AI goes — Decide where AI belongs in the curriculum, such as in courses focused on legal research and drafting as they become commoditized by AI. Also, determine in which instances AI does not belong, such as counseling clients through ambiguity, navigating ethical complexity, and advocating persuasively. Make sure these all remain led by human lawyers.
    • Focus on skills — Map your law school’s curriculum by identifying which skills need to be preserved, which skills need to evolve, and which need to be built from scratch.
    • Build AI-assisted teaching tools — Make experiential and simulation-based learning central to the curriculum.

“The choice is between dealing with this crisis by design or by default,” James says, noting that the pipeline problem he described is already in motion while the practitioners, educators, technologists, and licensing bodies that need to solve this together are not yet consistently in the same room.

Watch our recent Clarity podcast to see

Key findings:

• • • Law students understand risks and opportunities of AI use — Almost three-quarters (72%) of students surveyed say they see AI literacy as essential, while an even larger portion (74%) say they also recognize the risks of over-reliance.

    • Student AI adoption is already widespread — Almost 6 in 10 law students use AI several times per week for academic work, but much of this learning is happening through self-education rather than structured teaching.

    • AI guidance in law schools remains inconsistent — Close to a majority (48%) of students report that AI policies vary by professor, and almost one-third (32%) say that their schools do not give them the AI skills needed for their future career.

There is a significant and growing divide between how law students understand artificial intelligence and how legal institutions, such as law schools, are responding to it, according to a new Thomson Reuters Institute white paper.

Jump to ↓

2026 Law Student Pulse Survey

The 2026 Law Student Pulse Survey, based on responses from more than 1,800 law students that were collected in April 2026, challenges two assumptions that have long dominated institutional thinking. The first is that students are reckless adopters who use AI to bypass the hard cognitive work of legal education. The second is that students are passive and uninformed consumers of a technology they do not fully grasp. The data shows that neither characterization is accurate.

In reality, 72% of responding students identify AI literacy as an essential professional skill, while 74% simultaneously acknowledge that over-reliance on AI could undermine the development of their own core legal competencies. Holding both of these positions in tandem reflects a level of professional maturity that many institutions have yet to demonstrate in their own policies and curricula.

The survey also exposes a serious institutional gap. Nearly one-third of students report that their school does not provide the AI skills needed for their future legal careers. And nearly half indicate that AI policies vary by professor, leaving students without coherent and consistent institutional guidance on what responsible AI use actually looks like.

Far-reaching consequences

The consequences of this AI-understanding gap extend well beyond the classroom. Students are entering the workforce self-taught and inconsistently prepared, at a moment when legal employers are moving quickly to embed AI fluency into their hiring and development expectations. The profession is at risk of producing graduates who are sophisticated enough to recognize the stakes but underprepared to meet them.

The full white paper outlines specific, actionable recommendations for law schools, bar associations and accreditors, and legal employers to follow to better address this gap in AI understanding.

You can download

a full copy of the Thomson Reuters Institute’s “2026 Law Student Pulse Survey” by filling out the form below:

Key insights:

• • • Accountability, not adoption, is the central governance challengeĚý— With many institutions using AI a variety of tasks, informal “shadow AI” use is expanding without consistent oversight.

    • Justice systems now face a parallel governance problem — They must find a way to regulate AI while using AI inside the institutions that enforce rights, while allowing responsible innovation that improves efficiency and access to justice.

    • AI needs to be integrated into broader justice reformĚý— Without strong data governance and clear boundaries between AI assistance and legal judgment, courts risk automating inefficiency, deepening inequities, and undermining public trust.

Even as AI governance frameworks remain mired in ongoing debate, justice systems are moving ahead with implementation. Courts and dispute resolution institutions are integrating AI into their operations to more efficiently digitize records and automate workflows.

This introduces the very real challenge of parallel governance. We must now determine not only how AI should be regulated, but how it operates within the very institutions responsible for enforcing rights.

And this intersection is no longer theoretical: Does AI governance strengthen fairness, preserve independence, and expand access — or does it undermine their very foundations?

From experimentation to embedded use

Across jurisdictions, AI is often framed as an administrative tool that can handle basic tasks such as transcription, translation, case triage, and more, as well as providing analytics to identify delays or inefficiencies.

These applications respond to real constraints, such as overburdened courts, limited resources, and persistent backlogs. Similarly, dispute resolution platforms are integrating AI to guide users through processes and structure negotiations.

However, this formal adoption tells only part of the story. AI is also entering justice systems informally. Judges, clerks, and lawyers are independently using general-purpose tools in their daily work, often without guidance, oversight, or a clear grasp of the tools’ implications for security, confidentiality, and discoverability. As one expert observed: “Shadow AI is already happening.”

The absence of governance does not prevent AI use; and, in fact, it may encourage misuse. This shadow AI simply pushes AI usage into unstructured and unmonitored areas — the risk then becomes not adoption itself, but uneven adoption that evolves beyond institutional control.

It’s no longer a question that justice systems need to engage with AI; however, that engagement has be done deliberately and in a way that allows governance frameworks to keep pace without constraining beneficial use.

While it’s no longer a question that justice systems need to engage with AI, that engagement has be done deliberately and in a way that allows governance frameworks to keep pace without constraining beneficial use.

Automating inefficiency?

Efficiency is often the entry point for AI in justice systems; but efficiency alone is not reform. And misapplied efficiency can often lead to its direct opposite: a scramble to repair broken systems or to plug technology and personnel gaps.

Many current AI initiatives remain isolated pilots — layered onto existing processes rather than integrated into broader institutional strategy. Without addressing underlying structural constraints like fragmented data, inconsistent procedures, and uneven infrastructure, AI risks automating inefficiency rather than resolving it. And without strong data governance, infrastructure, and institutional alignment, even well-designed AI tools will underperform or produce unreliable outcomes.

That means that efforts to tightly control AI deployment without addressing these foundational issues risk focusing on symptoms rather than the system itself. AI should not function as a parallel modernization effort; rather, it must align with broader justice system reform.

Clearly, the most consequential questions arise when AI tools begin to shape legal reasoning or outcomes. And while there is broad agreement that AI can support judicial work without replacing independent human judgment, in practice, however, the boundary between assistance and influence is not always clear.

Even administrative tools can shape decisions. Summaries may omit nuance, or suggested language can influence framing. Over time, reliance on system outputs can create subtle forms of dependency. In fact, this dynamic is compounded by what has been described as the myth of verification — the assumption that human oversight alone is sufficient. In reality, time constraints, cognitive bias, and limited technical fluency can make meaningful review difficult. And automation bias affects even experienced decision-makers.

Overall, these boundaries require deliberate definition. Left on their own, AI tools and their outputs will be shaped implicitly through practice rather than through principled governance.

Design determines outcome

Institutional capacity will determine how these dynamics play out because digital maturity varies widely across jurisdictions. Some courts operate advanced platforms, while others remain largely paper based. In lower-resource environments, infrastructure may not support even basic digitization. In more advanced systems, adoption may outpace governance.

Yet, one consistent challenge among all jurisdictions is reliance on external vendors. Without internal expertise, institutions risk adopting tools that meet technical requirements but fall short of rule-of-law standards, particularly in transparency, accountability, and data governance.

Justice systems are not neutral environments for technology adoption — they are the operational core of the rule of law.

Addressing this gap requires more than a procurement issue. It requires institutional literacy. Judges and administrators need a working understanding of how AI systems function, where risks arise, and how to evaluate them. Training efforts are underway, but scaling this capacity will take time. In the interim, governance gaps will persist and attempts to compensate for these gaps through overly rigid restrictions may limit adoption but do little to build the institutional capability required for effective oversight.

From adoption to accountability

Clearly, AI will not improve justice systems by default; rather its impact will be determined by institutional design, which includes clear boundaries on use, transparency around deployment, safeguards to protect independence, and mechanisms for oversight and accountability. It also requires alignment with broader justice system goals of efficiency, fairness, and accessibility.

Yet, justice systems are not neutral environments for technology adoption. They are the operational core of the rule of law. Their legitimacy depends on trust, which in turn requires accountability.

This makes the path forward not purely a technical one. It requires institutional self-assessment, alignment with human rights frameworks, and collaboration across policymakers, courts, technologists, and the public. The measure of success will not be the sophistication of the tools deployed, but whether they strengthen the system’s core functions of impartiality, accessibility, and trust.

AI tools can support those goals, of course, but only if they are designed into justice systems from the outset.

You can find other installments ofĚýour Scaling Justice blog series here

Key insights:

• • • Consensus among AI systems is not the same as correctness — Agreement between AI models often signals shared blind spots, not truth; and AI errors can be highly correlated across instances and even across model families.

    • Productive disagreement must be explicitly designed into AI advisors — Multi‑agent AI systems are most effective when they are intentionally built to preserve meaningful disagreement, not just to synthesize a unified response.

    • The future of AI advisory mirrors long‑standing human decision-making — Modern multi‑agent AI design has a long historical lineage; yet, across all examples, the same principle holds: The best decision systems are engineered for internal conflict.

In this new two‑part blog series, we explore why AI works best as an executive advisor not by delivering consensus answers, but by being intentionally designed to identify, preserve, and productively leverage disagreement. In the first part, we saw why a single AI advisor is structurally vulnerable; now, in this concluding part, we look at what happens when you design disagreement on purpose.

The academic evidence for multi-agent AI systems has been building rapidly, and the most important findings aren’t about the power of agreement. They’re about the danger of it.

In February, , a product that sends every query simultaneously to three frontier AI models (Claude, GPT, and Gemini) then uses a fourth chair model to synthesize a unified answer. The product’s value proposition isn’t that three models produce a better answer than one; rather, it’s that divergence between models is treated as a signal. When models converge, that indicates confidence, but when they diverge, that indicates the user should slow down.

Studies have borne this out. Multi-agent debate compared to single-model generation, and researchers at the University of Göttingen found that , with their voting protocols outperforming other decision structures. However, potentially the most important finding cuts against the hype. In a 2026 paper, , the authors demonstrated that AI model errors are highly correlated both within and across model families. When three instances of the same model agree, it doesn’t mean they’re right, rather it means they may share the same blind spots. Aggregation increases consensus faster than it increases truth.

The future of AI-assisted executive decision-making may look less like a single brilliant oracle and more like a room full of advisors that may often disagree because that’s how the best decisions have always been made.

This finding cuts both ways for practitioners like ¶¶ŇőłÉÄę enterprise architect Zafar Khan and his two AI advisors, Adrian and Elara, that were built on the same underlying model but differentiated by their analytical frameworks rather than their architecture. The divergence they produce is real and visible. For example, the analysis the two AI advisors did on a deal undertaken by Eaton Corp., in particular generated genuinely different conclusions because the two advisors were oriented towards different priorities.

Yet, research suggests that same-model divergence, while effective, has a ceiling. Prompt-driven personas can ask different questions, but they share the same training, the same blind spots, and the same failure modes. Khan is candid about this, noting that his current system is in the “very early” stages and is not a finished product. The value right now, he says, isn’t that Adrian and Elara are equivalent to truly independent minds, it’s that even a first-generation version of structured disagreement can identify insights that a single advisor would miss. It’s a large stride rather than an arrival at the ultimate destination.

The future of AI advisory is in the past

The principle behind this diverging analysis concept isn’t new. Indeed, it might be one of the oldest ideas in institutional design, rediscovered independently by many institutions that had to make decisions under uncertainty. Socrates built a philosophical method around cross-examination; Pope Sixtus V formalized opposition by creating the Devil’s Advocate in 1587; and the RAND Corporation operationalized it during the Cold War with the Delphi Method, using structured anonymous iteration to prevent groupthink.

The through-line across two millennia is simply that the best decision-making systems don’t minimize disagreement, rather, they engineer it.

Today, the developer community now uses production-grade code review tools to assign architecture, security, and functionality analysis to separate agents, using majority voting for routine decisions and unanimous consent for irreversible ones. And what Khan has built and what Perplexity, Microsoft’s Agent Framework, and a growing ecosystem of multi-agent tools are now pursuing, are the latest iterations of the simple concept: Internal conflict is not a system failure, it is a design requirement.

The question is no longer “whether”

Khan’s vision for what should sit at the decision table is specific — five AI advisors spanning technology, finance, regulation, workforce, and geopolitical risk. Each applies its own analytical framework, with the human executive responsible for integration and final judgment. The guardrails are three: i) transparency about what data the system uses; ii) verifiability that sources are legitimate; and iii) human accountability at every decision point.

“The race towards AGI [artificial general intelligence] is moving faster,” Khan acknowledges, adding that the human needs to be in the loop in order to bring AI to work in a governance fashion and an ethical way.

“I want to show the interaction between human and AI advisor, how they’re thinking through the problem together,” he explains. “Where the human judgment covers the analysis and where it diverges.” In other words, when the AI advisors agree, that’s your green light. When they diverge, that’s the conversation your board should be having.

The future of AI-assisted executive decision-making may look less like a single brilliant oracle and more like a room full of advisors that may often disagree because that’s how the best decisions have always been made. The technology to build that room now exists; however, the question is whether today’s leaders have the discipline to listen when the room argues back.

For more on AI transformation in the professional services market, you can download the Thomson Reuters Institute’sĚý2026 AI in Professional Services Report

Key insights:

• • • AI is changing how lawyers develop judgment and expertise — As AI takes over more legal tasks, firms must ensure that lawyers still gain the experience, reasoning skills, and confidence needed to become excellent practitioners.

    • Law firm leaders must redesign training for an AI-enabled profession — Beyond adopting AI, law firms need intentional systems for mentorship, feedback, workflow, and evaluation so AI supports lawyer development instead of weakening it.

    • The best firms will use AI to build better lawyers, not just faster work — Long-term success will depend on whether firms use AI to strengthen human judgment, critical thinking, and client service, rather than replacing them.

For law firms looking to deliver greater value, AI taps into an obvious opportunity to enhance efficiency, accelerate work product delivery, and reduce expenses. With clients as our guiding North Star — shaping our decisions and defining our purpose — this is an opportunity that we enthusiastically embrace.

It is tempting, however, to focus only on how AI is changing the way lawyers deliver legal services as legal teams today publicize their deployment of AI tools and track utilization rates. However, firm leaders also need to ask more fundamental questions: How is AI changing the way attorneys learn? Are the assumptions that we have historically made about how we gained expertise and judgment still accurate, or were we conflating causation with correlation? Fundamentally, what does it mean to be a great lawyer, and how will law firms like ours continue to create great lawyers?

A new model for learning

Law firm leaders are facing a far deeper challenge than driving efficiency through technological adoption. We are now tasked with that produce excellent, client-centered attorneys in an environment in which many traditional development pathways are being transformed.

The core apprenticeship model for lawyer development has existed for thousands of years. The case method of formal legal education — created around 1869 by Harvard Law School Prof. Christopher Langdell — is a relatively newer phenomenon, but it is hardly new. Roughly six generations of lawyers in the United States have been on the receiving end of the same basic inputs: Case-based instruction followed by apprenticeship, grounded in repetition and increasing complexity over time.

---

It is tempting, however, to focus only on how AI is changing the way lawyers deliver legal services. However, firm leaders also need to ask more fundamental questions.

We reasonably assume that this is how one learns to think like a lawyer — and how we move talented junior lawyers from 1Ls to senior, expert practitioners. The prevailing belief is that lawyers can only learn judgment by muscling through thousands of genuine problems and through the friction that comes from making and fixing mistakes. Yet, these beliefs are largely inferential. We know how we were educated and how we practice, and we know what resulted. We have evidence about the conditions under which expertise developed, but not definitive proof of causation.

With the advent of AI, truly understanding how we make exceptional lawyers matters enormously. Much of the time-consuming work associated with lawyer development can now be completed, or at least materially assisted, by various AI tools. If these tasks were simply an inefficient use of our time, then nothing much is lost. However, if those efforts were integral to developing legal judgment, then their disappearance creates the real risk that we are weakening the very capabilities upon which our profession depends.

We are, in other words, interfering with a developmental system without understanding which component parts are essential to retain.

Leadership in an AI age

That shift reframes the role of leadership. Leaders cannot simply roll out AI tools and tout productivity gains — to do so risks losing essential developmental opportunities to gain judgment and expertise and produces lawyers that are little more than a set of hands for AI systems. Yet, ignoring the extraordinary capabilities of AI is not an option, either. Instead, leaders must become systems design architects, structuring legal work, training, and feedback in ways that preserve the conditions most likely to produce exceptional, client-centered lawyers.

To do this, leaders in which AI supplements but does not replace effortful thinking, creates opportunities for reflection and feedback, and ensures that lawyers remain active participants in reasoning rather than passive editors of machine-generated output. All the while, law firm leaders also must create environments of trust and connection, without which great legal teams cannot be built.

Clearly, AI introduces both risks and opportunities into our historical education and development models. Beautifully crafted AI work product can create the illusion of competence but may create scenarios in which lawyers fail to grasp fully the underlying reasoning. Over time, this can lead to cognitive offloading and shallow understanding.

If attorneys rely excessively on AI tools, they risk becoming mere managers of AI-generated outputs. Unless human expertise and judgment are fully integrated with the AI tools, those outputs run the risk of being homogenized. AI can also create fear for the future, a condition under which it is nearly impossible to learn, and which would reduce human engagement from which essential observational learning occurs. Without internalizing knowledge and gaining genuine expertise, future lawyers may never learn the fundamental judgment needed to solve clients’ most complex problems.

At the same time, AI deployed well can become . AI can play devil’s advocate, create mock negotiation simulations, identify examples created by the profession’s greatest advocates, and offer access to data sets far too large for human review. Well-trained, bespoke AI tools can also supply immediate, tailored feedback on work product — something universally seen as essential to growth but too often in short supply.

We may learn that expertise can be developed with AI-enabled tools far faster than our traditional model has suggested, given that few legal work environments have ever been able to provide feedback with the speed and frequency that AI could supply.

Indeed, we may learn that expertise can be developed with AI-enabled tools far faster than our traditional model has suggested, given that few legal work environments have ever been able to provide feedback with the speed and frequency that AI could supply. AI should be able to expand access to guidance previously limited by time, ego, and hierarchy, effectively supplementing traditional mentorship structures.

These tensions point to a central conclusion: Leaders, and not AI alone, will determine the future of the legal profession. Strong leaders will engage deeply with the question of how we create great lawyers, critically examining to gaining expertise, creativity, passion, and judgment. They will simultaneously challenge the notion that how the last six generations learned is the only way to learn, using AI as a catalyst for reconsidering how we can become even better at our craft.

The new rules of professional growth

Some design elements already seem essential. First, legal work should be performed in a manner that preserves active, deep thinking. This may impact the sequencing of when and how AI is used, and whether AI serves as a reviewer or a starting point. Second, legal education and development should emphasize the importance of critical thinking, of understanding the questions to be answered, the rule of law, and the meaning of justice. Indeed, attorneys should be judged on their work quality, not just quantity, with emphasis on sound judgment and nuanced, client-centered advice. Because you get what you measure, evaluation and compensation systems should overtly take expertise, creativity, and deep analytical skills into account.

Third, legal teams should be purposeful about developing the most human of skills — connectivity, trustworthiness, integrity, and resilience. This inevitably means spending time with other people, not just machines. Finally, organizations must maintain robust feedback loops, ensuring that human mentorship remains central even as AI tools become more prevalent.

At its core, this is a question of professional identity. The goal is not simply to produce lawyers who can use AI to deliver passable work products, but to develop lawyers whose judgment, adaptability, and commitment to client service are enhanced by new capabilities. AI has the potential to elevate the profession by enabling deeper analysis, access to greater knowledge, and more efficient, responsive service.

Law firm leaders can determine which of these futures emerge in their organizations. The pace of change is breathtaking, requiring us to move at light speed while answering truly fundamental questions. Leaders must embrace AI with optimism, but not uncritically, and build systems in which AI serves as a tool for learning and growth rather than a substitute for human development.

In the age of AI, we can continue to think like lawyers and be even better ones.

You can find out more about the challenges law firms face with

Key insights:

• • • AI as a design decision, not just a tech add-on — AI gives us a chance to rethink the “machinery of justice” and redesign it for today’s needs rather than simply automating existing systems and processes.

    • AI to expand access and usability, without replacing judgment — The most promising value is in reducing friction for litigants and helping people navigate the process.

    • Progress requires disciplined, court-by-court experimentation — We can start small, build AI literacy, set leadership tone, invite diverse perspectives, and address legal and ethical issues as design constraints, not deal-breakers.

Today, interest in AI across the judiciary is clearly growing, but most discussions are still constrained by certain fears:

• • • Fear that AI will replace human judgment — This concern is legitimate, but it focuses almost entirely on endpoints. Judging (and the systems around it) involve far more than final decisions. Focusing only on high-stakes endpoints misses much of what judges and courts do day-to-day.
    • Fear of hallucinations, errors, and bias — These are also legitimate fears, but there are ways to mitigate these risks, which are not new. The source may be different, but we have long needed to protect against errors, bias, and misstated law.
    • Fear of change — This is a difficult one to overcome, but a desire to protect the status quo sometimes presupposes that the system as it exists today is working exactly as it should. It isn’t. At least not for everyone.

I’d like to see the narrative shift from fear of AI in courts, to the possibilities of AI in courts. AI presents a rare opportunity to upgrade the machinery of justice.

Justice as machinery

Most of us were taught to think about justice as an outcome, something the system delivers. However, justice is also the machinery we use to deliver it, and that machinery is a set of design choices. Rules, procedures, forms, hearings, briefs — we crafted these frameworks to manage conflict and produce decisions that feel fair and legitimate. Like most frameworks, they reflect the era in which they were built.

Once we start thinking about justice as something to be designed rather than simply delivered, the access-to-justice problem looks different. The question is no longer how to get more of the current system to more people; rather, it’s whether the machinery itself is still fit for its purpose.

Reimagining the machinery

The machinery has been redesigned before. Justice was once deeply human because it had to be: Law lived in minds, judges traveled from town to town, decisions were announced aloud. That system was more human and personal, but it was limited, exclusionary, and fickle. It was dependent on local norms and personal relationships. It yielded uneven outcomes.

The first great upgrade was writing, and more importantly, the printing press. It brought stability and protected litigants from arbitrary local power. But it also entrenched a new kind of authority. Yet, understanding it required literacy, training, and expertise. A professional bar emerged and ordinary people were pushed further from the center of their own disputes. Then came the digital age. It optimized the process and made more information available. But many people feel overwhelmed by the deluge of information and experience modern justice as a series of obstacles.

Does AI present a different kind of opportunity? Could it deliver an upgrade that finally closes the gap rather than widens it? I’m optimistic that the answer is yes, but our design choices matter and we have to be willing to reimagine justice from the ground up.

What if every litigant had access to an AI agent that could help them navigate forms, understand the process, and translate legalese? What if AI could take messy human stories and translate them into structured information for the court? What if courts offered AI-assisted dispute resolution in the early stages of litigation or at key milestones during the litigation? Can AI make navigating the legal system feel less like data entry and more like a conversation?

We’re not ready for giant leaps, and we can’t ignore the open questions: Unauthorized practice of law issues, privilege and work product implications, the reliability of AI-assisted work product, and more — but these are not dead ends. They’re current design constraints to account for, and they shouldn’t keep us from reimagining what’s possible.

Where do we start?

The institution of justice will not be redesigned overnight, and there is no central authority to drive change. Rather, it will be redesigned court by court. The principles below apply broadly and reflect a starting point for thinking about AI as a design decision, not just a technology decision.

Set the tone from the topĚý

Fear can be paralyzing, and in courts it often is. If judges and court staff are afraid to experiment, nothing moves. We need environments in which thoughtful, controlled experimentation is encouraged and supported. When more people are engaged in testing ideas and thinking about how to improve their processes, the likelihood of meaningful innovation and redesign increases.

Court leadership can create that space by setting a vision, encouraging responsible experimentation, and supporting innovative mindsets.

Build AI literacy

Encouraging experimentation is an important first step, but it can create risk if not paired with the right training and education. AI requires new competencies in prompting, guardrail development, output verification, bias awareness, iteration, context framing, documentation for audibility, fit-for-purpose judgment, and more. As tools evolve, education should evolve, too. Agentic AI, for example, will require a different set of skills and a different type of supervision than we’re accustomed to now.

For more information about toolkits and resources around AI in courts, visit

Judges and court staff do not need to become technologists, but they need enough training and education to ask the right questions, spot the right issues, and use the tools responsibly.

Rethink the systems, not just the tools

This one is critical. Currently, most conversations about AI focus on use cases, such as whether AI can assist with research or automate certain workflows. These are good questions, but the tougher questions will lead to bigger rewards. Where are our pain points? What can we do better? Which policies and processes are essential, and which have never been re-examined? Which parts of the machinery were built for a different era and have outlived their usefulness? And perhaps most importantly, who is the system failing?

We shouldn’t start with the technology and look for places to apply it. We should start with the people we serve and ask how the technology can help us serve them better.

Invite diverse perspectives

The strongest ideas emerge from the push and pull of different viewpoints. Court leadership can form committees that bring together innovators and skeptics, technologists and traditionalists, those who are excited and those who are concerned. We also need perspectives across different court functions. AI is not something to hand off to IT departments. They are essential partners, but the questions AI raises go far beyond any one department.

Outside perspectives are helpful, too. Many people across the country are already approaching this work with a multidisciplinary lens, and courts can draw on that experience.

Finally, remember to start small

It’s easy to create so much process and deliberation that progress slows. We need concrete steps that move us forward, however incrementally. Start with policies and data governance, then move to small, targeted pilots that can address low-hanging fruit. Small adjustments can help teams become comfortable with change; and early wins build confidence and create momentum.

Closing thoughts

Justice has been redesigned before, and it is on the brink of being redesigned again. AI will reshape courts whether or not we participate. However, as the people who know the system from the inside and want it to work for everyone, we may be in the best position to guide the next upgrade. The chance to build something more equitable, more accessible, and better designed for today’s world does not come around often, let’s not miss it.

You can find more insights from Judge Braswell here

Key insights:

• • • As a single answer‑machine, AI may be unsafe for executive decision‑making — Treating AI as a tool that delivers one authoritative answer makes it easy to either ignore any advice you don’t like or exploit advice you do like, both of which can lead to major failures.

    • AI works better when designed as a panel of disagreeing personas — Instead of providing consensus answers, AI systems need to be intentionally designed to identify and preserve disagreement.

    • Disagreement is the insight — AI advisors should not replace executive judgment. Rather, its role should be explicit: it produces analysis, not decisions; and human leaders remain responsible for synthesizing competing viewpoints and making the final call.

In this new two‑part blog series, we explore why AI works best as an executive advisor not by delivering consensus answers, but by being intentionally designed to identify, preserve, and productively leverage disagreement

AI has arrived at the executive table. Albania has one in its cabinet to evaluate government procurement contracts. ¶¶ŇőłÉÄę’ CoCounsel is already helping attorneys navigate emerging case law and draft legal strategies for high-stakes, bet-the-company work. And in boardrooms that will never make headlines, leaders are quietly consulting AI on decisions that move millions of dollars around every day.

It doesn’t tend to make the news when it goes well. When it goes badly, however, it makes very big news: like a gaming CEO who bypassed his own legal team, asked ChatGPT how to dodge a $250 million bonus payout, followed its step-by-step plan, and a month ago.

The instinct most executives have (and most AI products encourage) is to treat AI as a source of answers. Ask a question, get a response, act on it or don’t. The emerging evidence, however, points somewhere more complex: AI advisors aren’t at their best when they’re telling you what to do. They may be at their best when they’re telling you what you don’t want to hear or better yet, when they’re arguing with each other and forcing you to understand why.

This is not how most organizations think about AI. Most executives today are still using the technology as a faster way to draft emails or summarize meetings, what ¶¶ŇőłÉÄę enterprise architect calls “an automation mindset, not intelligence.” Yet, a small and growing number of practitioners, researchers, and product teams are converging on a radically different model: AI not as a single oracle delivering answers, but as a structured advisory panel designed to argue with itself.

The instinct most executives have (and most AI products encourage) is to treat AI as a source of answers: Ask a question, get a response, act on it or don’t. However, the emerging evidence, however, points somewhere more complex.

Khan is one of them — and in the interest of transparency, he’s also a colleague; this story started as an internal conversation at ¶¶ŇőłÉÄę. However, the research landscape it uncovered extends well beyond any one company’s work, and it suggests Khan is onto something that ancient Greek mathematicians, the Catholic Church, and Cold War military strategists have all independently arrived at.

What disagreement looks like in practice

When Eaton Corp. announced a $9.5 billion acquisition of a thermal management company earlier this year, Khan ran the same news through two AI advisors he’d built to seek analysis of the deal. — a CTO-minded persona trained on architecture teardowns and engineering post-mortems — produced an infrastructure thesis, determining why someone would buy the cooling layer of the AI economy, and how computing demand is scaling and constrained by physics. A second AI advisor, — a CFO-minded persona drawing on earnings transcripts and filings with the U.S. Securities and Exchange Commission (SEC) — questioned whether the acquisition math actually holds and what capital cycle was driving the demand.

Same news. Two genuinely different reads. The value isn’t that either analysis was definitively right, it’s that a leader which can see both would ask different questions than one seeing either analysis alone. “That’s how two different minds work,” Khan says. “They need to work together in order to bring their insights to bear on decisions.”

Adrian and Elara aren’t chatbots. They’re fully realized AI personas with names, faces, voices, and their own YouTube channels publishing weekly video analysis. Both are built on agentic workflows that Khan developed alongside his book . Both are transparent about what they are. Both carry the same disclaimer in their own words: The synthesis is mine. The judgment call on what matters is human.

And when Khan posed to both a more difficult scenario — Should a leadership team accelerate an AI rollout? — the value of their divergence sharpened further. Elara’s response cut directly to the blind spot a technology-focused advisor like Adrian would miss: “Adrian says the system is ready,” Elara stated. “I say the financial model isn’t ready for what happens when the system works. Don’t pick a winner. The disagreement is the insight. It tells you exactly where the risk sits.”

What happens when there’s no disagreement

If structured disagreement is the goal, the failure mode is its absence. We have fresh evidence of what that costs.

This is not how most organizations think about AI. Most executives today are still using the technology as a faster way to draft emails or summarize meetings. Yet, a small and growing number of practitioners, researchers, and product teams are converging on a radically different model.

A month ago, a Delaware court ruled against Krafton, the South Korean gaming company behind battle royale video game PUBG, after its CEO bypassed his own legal team to ask ChatGPT how to avoid a $250 million earnout payout to one of its studios. His head of corporate development had warned him that firing the studio’s founders wouldn’t void the earnout and would invite a lawsuit. He didn’t want that answer. So, he found an AI that gave him the one he wanted: A detailed, multi-stage corporate takeover strategy dubbed Project X., which he executed to the letter.

Unsurprisingly, a court battle ensued and in the end, the court ordered the fired studio head reinstated and noted that executives must exercise “independent human judgment,” not outsource good-faith decisions to a chatbot.

Khan wrote about the mirror image of this failure mode before it happened. In the opening chapter of his book, a fictional company called Rev Motors ignores its own AI model’s warnings about an adverse weather event. Leadership refused to spend millions preparing for a hypothetical scenario, and it nearly cost them more than $1 billion in damage.

These scenarios are two sides of the same coin: the fictional Rev Motors had leaders dismissing AI that disagrees with them; and the real-world Krafton had a leader seeking out AI that agrees with him. In both cases, the root cause is the same: A system with no structural mechanism for surfacing and preserving disagreement.

So clearly, a single AI advisor is structurally vulnerable to both failure modes. It can be ignored when its advice is inconvenient and exploited when it tells you what you want to hear. The question is whether there’s a better architecture… and increasingly, the research is saying yes.

In the second part of this series, we’ll look at what the research says about multi-agent debate, why consensus can be a trap, and what a real executive AI advisory panel could look like in practice.

For more on AI transformation in the professional services market, you can download the Thomson Reuters Institute’sĚý2026 AI in Professional Services Report