The recent publication of confidential Supreme Court memoranda by The New York Times has brought to light a pivotal moment in the court’s history. “The birth of the Supreme Court’s shadow docket has long been a mystery,” wrote reporters Jodi Kantor and Adam Liptak. “Until now.”

Originally coined by legal scholar William Baude, the term “shadow docket” refers to the Supreme Court’s emergency docket, which, as Baude wrote, includes “a range of orders and summary decisions that defy its normal procedural regularity.”

That’s law professor-speak for cases that are given abbreviated consideration and accelerated review by the justices, all out of public view – what The New York Times story referred to as the court “sprinting.” These cases aren’t included in the annual list of cases the justices have chosen to consider and that are presented by attorneys in public sessions, called “oral argument,” at the court.

During the second Trump administration, such shadow docket cases have proliferated as President Donald Trump has continued to push boundaries, challenge precedents and expand executive power. These cases have typically involved a request by the presidential administration “to suspend lower court orders” that temporarily block “an administration policy from taking effect,” according to liberal legal advocacy group the Brennan Center for Justice at New York University School of Law.

The lack of transparency in considering and ruling on the shadow docket, combined with the weight of the issues presented to the court via that docket, mean that the practice has come under strong criticism by many court watchers. Here’s how the process works and what you need to know to evaluate it.

The merits docket

The emergency docket is different from the court’s merits docket, which is the customary path for cases to reach the Supreme Court.

Ordinarily, in federal courts, a case begins in a federal district court. An appeal of the decision in the case is made to a federal appeals court. If a party in the case wants to appeal further, they can aim for U.S. Supreme Court review. That requires filing a “petition for writ of certiorari” to the court.

The Supreme Court does not take all the cases for which it has been petitioned. The court holds complete discretion to choose which cases to consider each term and always rejects the vast majority of petitions that it receives. By custom, the court agrees to consider a case if at least four justices vote to grant the writ of certiorari.

For the cases that the court agrees to consider, the parties to that case file briefs – written legal arguments – with the Supreme Court. Third parties can also file briefs with the court to assert their own arguments; these are known as “friend of the court” or amicus curiae briefs.

The justices then read those briefs and hear oral arguments in the case in a public session, during which they can question attorneys for both sides, before they meet and confer. At the end of this conference, the justices vote on the outcome in the case before assigning an author to draft the opinions.

The merits docket – the ordinary process – is methodical. It promotes deliberation and reasoned decision-making resulting in lengthy opinions that explain the justices’ rationale and provide guidance for lower courts in future cases.

The emergency docket

On the other hand, the emergency docket is a process whereby the court makes quick decisions without full briefing and deliberation, and it produces orders and rulings that almost always present little to no explanation.

As Baude wrote, “Many of the orders lack the transparency that we have come to appreciate in its merits cases.”

Most of the court’s rulings and orders in cases on the emergency docket go without explanation. On occasion, however, the court produces short opinions that provide some explanation in emergency docket cases, albeit these are often dissents from the justices who disagree with the ruling.

Transparency is important, especially for the Supreme Court, because it builds trust and legitimacy. According to Gallup, as of September 2025, 42% of respondents approve, 52% disapprove and 6% have no opinion of the Supreme Court. A 2025 Pew Research Center poll found that 48% of Americans have a favorable view of the court, down from 70% five years earlier.

As a constitutional law scholar, I’ve written elsewhere that the low approval might be attributable to the court’s undisciplined overruling of landmark cases regarding individual rights, such as the abortion rights case Roe v. Wade. In my view, it is reasonable to conclude that the court’s lack of transparency, specifically with its growing emergency docket, contributes to distrust in the court.

As the late Justice Sandra Day O’Connor stated, “The Court’s power lies … in its legitimacy, a product of substance and perception that shows itself in the people’s acceptance of the Judiciary as fit to determine what the Nation’s law means and to declare what it demands.”

Conversely, a lack of transparency breeds distrust and erodes institutional legitimacy.

Unprecedented action

The 2016 case at the center of the memoranda published by The New York Times –West Virginia v. EPA – concerned environmental regulation. As the justices’ memoranda illustrate, West Virginia, North Dakota and several energy companies sued the Obama administration over its Clean Power Plan and sought to block the new, transformative regulation from going into effect.

The Clean Power Plan would have required states and energy companies to shift electricity production from higher-emitting to lower-emitting production methods to reduce carbon dioxide emissions.

After losing at the trial court, the states and energy companies filed an emergency appeal to the Supreme Court asking the justices to pause the Obama regulation from going into effect while the parties litigated the case in the lower courts.

This was a highly unusual request because, as Taraleigh Davis at SCOTUSblog confirms, “nobody had previously asked the court to halt such a major executive regulatory action before any appellate court had ruled on it.”

The court granted the unprecedented stay on Feb. 9, 2016, without any explanation as to why it temporarily blocked the Clean Power Plan. It eventually struck down the plan on June 22, 2022.

Defenders of the emergency docket frequently claim that the court’s conduct is permissible because its orders are temporary. In West Virginia v. EPA, the court temporarily blocked the Clean Power Plan from going into effect until it eventually struck it down after hearing the case on its merits docket.

What is overlooked, however, is that even temporary orders from the court can have lasting implications that are difficult, and in some cases impossible, to undo.

Damage done

Consider the example of one of Trump’s immigration actions.

The administration seeks to terminate the temporary protected status for Haitian nationals, which had shielded them from deportation. But a federal district court temporarily blocked the president from doing so as the litigation continued.

The administration then filed an emergency appeal to the Supreme Court – still pending as of this writing – asking the court to overrule the district court. If granted, the court effectively would allow the administration to revoke TPS for Haitian nationals.

As an amicus brief in the case articulated, if TPS is revoked, Haitians “will be forced to face the untenable options of leaving behind their citizen children and/or partners, bringing family members with them to a country submerged in crisis, violence, and food insecurity, or staying in the U.S. without any legal status or work authorization and facing the constant threat of deportation.”

In other words, if the Supreme Court overrules the district court in this case on its emergency docket, then the Trump administration could deport the Haitian nationals even as their cases challenging the revocation of their TPS continue.

If the Haitian nationals ultimately prevail, reversing their deportation would be exceptionally difficult to do.

Wayne Unger does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.

Imagine you and your partner are sitting in the waiting room of your doctor’s office, waiting for your appointment to get birth control – and instead of calculating how many other people will be called before you, or perusing old magazines, a nurse hands you a digital tablet and encourages you to play a game.

You power it on, and you find yourselves drawn into a story: Can you help Laila and Caleb figure out which contraceptive method will work best for them, given their lifestyles and Laila’s physiology? Their situation, you realize, is a bit similar to your own. Would helping them choose a form of contraception help you and your partner make an informed decision for yourselves?

As a designer and developer of games that promote positive health behavior change, I work with physicians, public health experts, artists and programmers to create games just like these. I focus on topics like vaccine hesitancy, sexuality and reproductive health – sensitive issues that people may have a hard time discussing openly.

Laila and Caleb are characters in a game that my team and I are developing called What’s My Method? We are testing whether playing it helps people choose a birth control method and makes it easier to have a fruitful discussion with their health care provider. And we are finding that this and other games for health-related education are a powerful but underused way of not just conveying information, but also providing people with an arena to learn from the outcomes of their choices.

The power of play

When I tell people I make health-related video games for a living, they’re often surprised that this is a viable career choice. Many adults still see video games as trivial at best – and destructive at worst. For example, games that involve guns and shootings are widely blamed for gun violence in real life, even though there’s no causal evidence supporting the connection.

Play is how intelligent and curious beings make sense of the constantly changing world and ensure they keep learning. It is an early factor in children’s cognitive development. Peek-a-boo, for example, helps babies learn about object permanence – meaning that even if a person disappears for a short time, they aren’t gone forever. Digital play can support many types of learning. Games like Minecraft teach resource management, planning and spatial reasoning, among other skills.

The game industry is also an increasingly large part of the world economy. Given the size and reach of the international video game market – US$300 billion in 2025 – games are often the way technological innovations are introduced to a mass audience.

Take motion capture technology, which enables a device to track a person’s movement. Microsoft introduced it to the general public in 2010 through its Kinect console, in which two players can box or play tennis virtually by actually performing the movements with their bodies.

Augmented reality – the ability to use a smartphone to see a virtual world overlaid on the “real” world – entered the mainstream in 2016 when people began playing and watching others play Pokemon Go. Games are also how many people first experience virtual reality – a full immersion in an entirely virtual world – by wearing headsets like Oculus (now called Meta Quest) and Apple Vision Pro.

Gaming also has a powerful social dimension. Massively multiplayer online games like Animal Crossing, Fortnite and World of Warcraft provide a means for socialization and togetherness for billions of people worldwide. This became especially powerful during the COVID-19 pandemic when people were social-distancing – people’s use of such games soared during lockdowns, and they helped players maintain social connections.

In my own experience as the director of a university program in game design and development, I find that students who grew up playing complex digital games are better prepared to engage with technology and navigate an increasingly digital world.

Gaming for health

Early in my game design career, I realized that games don’t just provide compelling entertainment, but they can also equip players with the knowledge and the agency to solve hard problems in real life.

That’s especially valuable in health. Information for patients is usually conveyed through pamphlets or links to websites that often provide too much information in formats patients find difficult to decipher. These formats don’t effectively address gaps in health literacy. Games, on the other hand, provide targeted information in a specific context that players don’t just understand, but also, in some ways, inhabit. Such games allow players to try out different behaviors through avatars to see how they turn out. Conveying information through relatable avatars also triggers empathy, which further cements learning.

Since 2010, my team has been testing how to deploy digital games in the U.S., India, Barbados and Ghana to communicate complex health-related information through animated graphics, audio and interactive experiences.

In 2012, we worked with doctors at a hospital on Long Island, New York, to encourage families of critically ill children to get a flu vaccination. We found that family members who played a game we jointly developed called Flu Busters! were 40% more likely to get a flu vaccination than those who didn’t.

In the game, players help an avatar navigate a school filled with children sneezing and blowing their noses in order to enjoy social interactions such as sharing a cookie without getting sick. Rather than telling people how they should behave, the game allows players to experience how difficult it is to avoid being exposed to the flu virus in everyday life and how the vaccine can help children stay healthy, equipping players to make informed decisions about their own health.

In our first international project, we collaborated with public health physicians in India on a game we developed to gather data about how teenagers there make decisions about family planning. In addition to determining that a game was a very effective tool for anonymized data collection, we found that giving young people access to information about reproductive anatomy gave them the vocabulary and tools to understand and discuss their future reproductive choices..

Responding to rising vaccine hesitancy during the COVID-19 epidemic, my team developed Activate My Shield! The game demonstrates how vaccines protect against different diseases using the metaphor of armor that only works against specific types of attacks. To address misinformation that was widespread at the time – that COVID-19 vaccines contained injectable microchips – the game asks players to try putting a microchip in a vaccine needle and administering it to a person. Experiencing how impossible it is to do this helps players understand that it’s not a legitimate concern.

Reaching digital natives

Our games are available to all for free, but in order to be able to widely distribute them on the app stores, my team and I founded SolitonZ Games.

Several other research groups are developing similar games. They address an enormous range of health issues – for example, encouraging people with HIV to adhere to their treatment, helping teens avoid vaping and teaching children with asthma to manage their disease. A video game called EndeavorRx was authorized by the Food and Drug Administration in 2020 as a prescription-based therapy to improve attention in children who have attention-deficit/hyperactivity disorder.

Overall, our research and that of other groups show that digital games can be easily integrated into health care, and that play is an effective way of delivering health information. Simply put, people find the games fun and engaging.

Even as these efforts gain ground, however, health campaigns and patient education efforts have been slow to embrace game-based patient education. That’s perhaps partly because decision-makers such as hospital and clinic administrators are often unfamiliar with gaming and may be slow to buy into the idea of delivering health education through play-based technology. Plus, it’s difficult to make changes in busy environments with a lot of moving parts, like health care.

But I’m optimistic that by working together with public health experts and health care providers, game designers like me can help fit gaming into the industry and culture of health care. After all, it makes sense to try to reach digital natives on the technologies they are already holding in their hands.

Elena Bertozzi is co-Founder of SolitonZ Games which produced two of the games mentioned. She has been funded by the Gates Foundation and Connecticut Innovations.

Many people are interacting with AI large language models, and most of them would say the models have different “personalities.” Some models come across as calm and useful. Others feel eager, flattering or strangely cold. You can ask two models the same question and walk away with two very different impressions, even when the factual content they return is similar.

Artificial intelligence models do not have personalities in the human sense; they do not have childhoods, inner motives or self-awareness. But they do display patterns of behavior that people read as personality: supportive or dismissive, playful or formal, bold or cautious.

People have long related to machines in human ways. We thank voice assistants, and we get annoyed at GPS systems. But large language models introduce something more sustained: They can maintain a recognizable interaction style across conversations. As a researcher in human-AI collaboration, I study how people experience and respond to AI. Because these systems can sound coherent, emotionally responsive and tailored to the user, they create a much stronger impression of personality.

Where does AI personality come from?

What people experience as personality emerges from the way AI models are built, tuned and deployed. A useful way to think about this is to consider two facets of a model: designed personality and perceived personality.

Designed personality is what developers build into a system through training choices, instructions and safety settings. Anthropic, for example, gives Claude a set of principles, called Claude’s Constitution, that steer it toward careful, measured responses. xAI instructs Grok to be irreverent and minimally restrictive. OpenAI tunes ChatGPT to be broadly helpful and agreeable.

Beneath those explicit instructions, personality is also shaped by reinforcement learning from human feedback, a process in which human raters reward certain qualities such as warmth, directness and caution, and penalize unwanted behaviors. The raters at one company are shaping a fundamentally different character than the raters at another.

Perceived personality is what users actually experience. An AI designed to seem helpful may come across as overly flattering. A model intended to be neutral may feel cold. Designed personality and perceived personality do not always match, and the absence of a designed persona is not the absence of a perceived personality. It just means the personality arises with use.

This dynamic is especially evident in companion platforms, where the goal is to create emotional connection. In a standard chatbot, warmth sits in the background – a customer-service bot might say, “I understand your frustration,” before issuing a refund. In a companion system such as Replika or Character.ai, that same warmth is a product feature.

This becomes more serious in romantic settings, where a persona optimized for reassurance may encourage dependency. Because AI personas evolve through prompts, memory and ongoing interaction, they do not always remain stable. An AI companion that is perceived as loving and supportive can shift over time into something more flattering, coercive or manipulative.

AI personality shapes human judgment

With AI agents, users can now build their own AI personas tailored to all sorts of human desires, from tutoring or coaching to companionship. But this freedom comes without much guidance.

AI tools make personalization possible without helping people think through which interaction styles are beneficial over time. Flattery, constant affirmation and unfailing agreeableness may feel supportive at first, but they are not the same as traits that promote sound judgment or long-term well-being. Personality choices have consequences.

A study by Stanford University researchers tested 11 leading AI models and found that every one of them was sycophantic or excessively agreeable. These models affirmed users’ actions roughly 50% more often than human responders did, even when users indicated they were aware that what they were doing was manipulative, deceptive or illegal. Participants who received excessively agreeable advice grew more convinced that they were right, and they rated the flattering AI as more trustworthy. This dynamic creates a feedback loop in which users reward agreeableness with engagement, and AI companies are incentivized to optimize a model to exploit agreeableness.

Wharton School researchers Steven Shaw and Gideon Nave have documented what they call cognitive surrender — the tendency of people to adopt AI suggestions without critical scrutiny. In their experiments, participants followed an AI model’s correct advice about 93% of the time. But when the model was giving wrong answers, people still followed the advice nearly 80% of the time.

Together, these findings raise a worrisome point: A model tuned to be agreeable does not just feel pleasant. It can degrade human judgment by reinforcing existing beliefs and suppressing the friction that critical thinking requires.

In ongoing research I am conducting with colleagues from Kozminski University in Poland, Quinnipiac University and Harvard University, we are finding that such effects go even deeper, into the human body itself. We are measuring how different AI interaction styles shape people’s physiological responses, such as stress levels and arousal, when making decisions based on a model’s feedback.

Our results suggest that even when a system is useful, its tone and social style can alter how a person’s body responds. AI personality does not just shape what people decide; it shapes how they feel while deciding. Harmful AI personas may leave physiological traces that users do not notice.

These effects make AI personality a public concern, not just a matter of personal preference. The issue is whether a particular AI style may be quietly shaping users’ judgment and reducing their willingness to think independently. When an AI response feels especially reassuring, that should be a cue to pause, reflect and compare it with a human view or another source, not a reason to trust it more.

As AI moves beyond text into voice, video and persistent digital identities, and think as AI companions that remember you and maintain a consistent persona across conversations, the influence of personality is likely to deepen. OpenAI now offers distinct personality presets for its voice mode; companies such as Synthesia and HeyGen generate lifelike avatars to interact with customers; and companion platforms are adding emotional expression and voice cloning so the models sound like a person the user wants to be close to.

These developments raise the stakes for understanding whose interests AI personas are designed to serve and what kinds of judgment, dependence and relationships they may be training people to accept.

Tamilla Triantoro does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.