Stock

Patrick G. Eddington

Late on Friday, February 7, President Trump issued an executive order titled Protecting Second Amendment Rights. The order directs newly confirmed Attorney General Pam Bondi to 

…examine all orders, regulations, guidance, plans, international agreements, and other actions of executive departments and agencies (agencies) to assess any ongoing infringements of the Second Amendment rights of our citizens, and present a proposed plan of action to the President, through the Domestic Policy Advisor, to protect the Second Amendment rights of all Americans.

The core language of the order is aimed squarely at Biden-era policies, specifically reviews of presidential and agency actions spanning January 2021 through January 2025. However, Section 2 (v) directly addresses past, and by extension potential future, executive branch positions in court cases involving gun rights:

The positions taken by the United States in any and all ongoing and potential litigation that affects or could affect the ability of Americans to exercise their Second Amendment rights…

Thus far this term, the Supreme Court has declined to take up cases involving state-level bans in Maryland and Illinois on so-called “assault weapons,” a pejorative, inaccurate term used by anti-Second Amendment activists to equate semi-automatic modern sporting rifles with fully automatic military and law enforcement rifles like the M4 Carbine or it’s replacement and the Army’s new battle rifle, the Sig Sauer XM5. 

The one major case currently before the high court involves Mexican government allegations of gun trafficking by American gun manufacturer Smith & Wesson to drug cartels in Mexico. That case is scheduled for oral arguments on March 4. Whether Trump’s order will now prompt administration action in connection with that case remains to be seen. 

Not included in Trump’s order was any language directing Bondi to conduct a review of existing federal laws affecting Second Amendment rights, a curious omission given the otherwise sweeping nature of the order.

John F. Early

The Census Bureau has filed a Federal Register notice of its intent to change and expand its coding of race and ethnicity in the 2030 decennial census and the ongoing American Community Survey, with any comments due by February 18, 2025.[1] Census claims that it is “not seeking feedback on how the US Office of Management and Budget defined race/​ethnicity categories through Statistical Policy Directive No. 15.” That, of course, is a specious exclusion because a failing in SPD 15 may appear in the Census proposal and, as such, must be addressed irrespective of its origin. As this paper will show, the Census proposal actually conflicts with some provisions of SPD 15 and ignores opportunities in the directive to correct some existing failures, which requires reference to that source.

There are three classes of fatal flaws in the Census proposal that should motivate it to abandon this proposal in its entirety. They are as follows:

Adopting this proposal would violate the principles of scientific objectivity that should be the foundation of all Census work.
It is unconstitutional.
It is wrong on multiple significant historical and scientific facts.

Each of these is discussed in detail below.

Violates Census Foundation of Scientific Objectivity

Historically, the Census Bureau has observed the foundational principle of scientific objectivity— counting and reporting the things that one can observe—the number of people, the amount of manufacturing production, household income, etc. The categories of race and ethnicity fail that test of objectivity. The SPD 15 from the Office of Management and Budget (OMB) says, “The categories in these standards are understood to be sociopolitical constructs and are not an attempt to define race and ethnicity biologically or genetically.”[2] In other words, the whole scheme is a collection of politicians’ and bureaucrats’ personal opinions, not objective scientific facts.

The subjectivity of the scheme is further demonstrated by the fact that race and ethnicity are classified by the personal opinions of the respondents as to their racial origins—a feature that has been observed to be highly inaccurate and even intentionally misrepresented. Unlike other features of Census surveys such as employment or income, there is no objective way to evaluate these responses and correct for nonresponse or misreporting because the classification is purely subjective in the first place. More examples of how this proposal violates the canons of science and objectivity are discussed later.

Census should reassert its long but recently weakened tradition of scientific objectivity and refuse to follow these subjective, inherently unverifiable classification schemes. It has the authority to regain its scientific respect by stopping all racial classification because (1) SPD 15 specifically says, “The standards do not require any agency or program to collect race and ethnicity data,”[3] and (2) the Census authorizing legislation has no such requirement.[4]

Unconstitutional

The decennial census is the only data project that is explicitly authorized in the United States Constitution. Article 1, Section 2 requires:

Representatives and direct Taxes shall be apportioned among the several States which may be included within this Union, according to their respective Numbers, which shall be determined by adding to the whole Number of free Persons, including those bound to Service for a Term of Years, and excluding Indians not taxed, three fifths of all other Persons. The actual Enumeration shall be made within three Years after the first Meeting of the Congress of the United States, and within every subsequent Term of ten Years, in such Manner as they shall by Law direct.

With the elimination of slavery, the Fourteenth Amendment changed that language to:

Representatives shall be apportioned among the several States according to their respective numbers, counting the whole number of persons in each State, excluding Indians not taxed.

The 1924 Indian Citizenship Act effectively eliminated the class of “Indians not taxed.” As a result, while the original Constitution laid the basis for collecting census data for African Americans (although not all of them were enslaved) and Indians, the Fourteenth Amendment and the subsequent congressional act removed any justification whatsoever for any racial or ethnic distinctions in the collection of the Census data.

The “equal protection of the laws” provision of the Fourteenth Amendment has become foundational to the clear and strong prohibition of any discrimination by the government based on ethnicity, race, or national origin. As Chief Justice Roberts has famously said, “The way to stop discrimination on the basis of race is to stop discriminating on the basis of race.”[5] In issuing SPD 15, OPM seems to be aware of this injunction and says of the racial and ethnic classifications: “They are not to be used as determinants of eligibility for participation in any Federal program.”[6] But, SPD 15 is schizophrenic because later it says: [7]

“These standards shall be used for all Federally sponsored statistical data collections that include data on race and ethnicity.”
“These standards shall be used for all Federal administrative reporting or record keeping requirements that include data on race and ethnicity.”
“These standards must be used by all Federal agencies for civil rights and other compliance reporting from the public and private sectors and all levels of government.”

So after saying these standards cannot be used to determine who is rewarded and who is punished, SPD 15 incoherently claims that if you nonetheless use any racial data to determine reward and punishment, then it must use these categories.

Such requirements directly support administrative and even criminal civil-rights proceedings by the government that will use these data to make determinations of some sort to favor or disadvantage individuals based on their race. That is unconstitutional and antithetical to the foundational beliefs and principles that underly our society and governance. 

The punitive intent and potential for these classifications are made even more objectionable by Census advertising prominently to respondents for data collection that includes race, “you are legally obligated to answer all the questions, as accurately as you can. The relevant laws are Title 18 U.S.C Section 3571 and Section 3559, which amends Title 13 U.S.C. Section 221.”[8] The consequences are staggering. Government can punish you for answering a subjective question if the answer does not comply with its subjective categories.

Wrong on multiple significant historical and scientific facts

The racial classification scheme begins with six “mandatory minimum” categories of American Indian or Alaska Native, Asian, Black or African American, Hispanic or Latino, Middle Eastern or North African, Native Hawaiian or Pacific Islander, and White. Each of these is defined with the formula: “X are individuals with origins in any of the original peoples of Z.” The group named X is all those descended from the “original peoples” that lived in Z.

This racial formula is entirely wrong and inconsistent with history and biology. In one sense, the “original peoples” for every one of us human beings begins with the early Homo sapiens, the starting point of all “original people.” We all began from the same Homo sapiens (or according to some formulations the same earlier Homo habilis), all of which originated in the broad African Savanah in the vicinity of Olduvai Gorge. That most original of all original peoples cannot be found anywhere in the proposed classification. From those earliest humans, we spread out across the planet for the next 200,000 years. Invariably the “original peoples” on almost any square mile of earth were later replaced by another set of people as the huge global migration spread—sometimes motivated by weather or overcrowding, but often by slaughter, expulsion, or subjugation.

The Census classification scheme seems unaware of human history, even modern history in this great age of literacy, printing, and electronic communication. The Census lists “Turks” as among the “original peoples” of Europe, and hence “White,” although only the small trans-Bosporus portion of Turkey lies within the usual definition of Europe. The United Nations, Turkey itself, and most reference works list Turkey as part of Asia, Asia Minor, or the Middle East, which areas constitute the vast majority of Turkey’s land mass and contain the capital Ankara as well as most of the population. 

But as silly as that designation of Turks as European may be, the claim that they are the “original people” of that land is pure fiction. Until the mid-fifteenth century, the “original peoples” of that piece of land included, among others, the Greeks, Armenians, and Kurds. The Turks moved in from the eastern-central Asian steppes and systematically replaced these populations of long-standing via slaughter, deportation, and suppression, which continued well into the 20th century and even to this day.

Shortly before the Turk destruction and replacement of the Asia Minor populations and still well within the literate historically documented times beginning in the seventh century, Arabs moved out of the Arabian Peninsula and across North Africa, displacing the predecessor populations of Berbers (also called Imazighen or Amazigh), Sahrawis, and other populations. Yet the Census lists Arabs, Moroccans, Algerians, Tunisians, and Libyans as “original peoples,” despite their recent arrival and displacement of the Berbers and their ken. The Census classification does list the Berbers independently but shoves the Sahrawis into the category “Middle Eastern or North African, Not Elsewhere Classified.”

The entry in SPD 15 for “White” would be hilarious if it were not so ignorant. It reads: “Individuals with origins in any of the original peoples of Europe, including, for example, English .…” English is the first “original people” listed in “White.” Julius Caesar in De Bello Gallico in 55 BCE described the people of what we now call England as “Britanni,” not English. He also described some of them as having come from Denmark. The term Briton endured in the local languages as their identity even after the Romans largely withdrew in the fifth century.

Genetic evidence shows that most “English” folks have strong strains derived from ancient stock that we now call “Celtic,” which exists in identifiably different strains in Ireland, Scotland, Wales, and England. The contemporaneous but somewhat separate Picts are also believed to derive from the same Paleo-Celtic stock. These Celtic peoples moved to what is now the British Isles about 12,000 years ago as the last great ice age retreated, replacing the Paleolithic people who had lived there at least 25,000 years ago until the extreme ice-age cold drove them out. Genetic studies show that the Celtic origins trace back to migrations from Iberia up the coast to Britania. (So, should one classify “English” as Hispanic since SPD 15 defines Hispanic as “of … Spanish culture or origin,” and “Iberia” is the geographic area of Spain?)

As the Roman legions slowly withdrew in the fifth century, they left behind gene pools from all over the Roman Empire, including the Middle East and Africa. Those were quickly layered over with additional Germanic invasions—the Angles (from whose name “England” was derived), Saxons, and Jutes. Those were followed by the Vikings—also separately identified as Danes, Norwegians, or Swedes—and the Norman French in 1066. Britania is not by any means unique. This vast sequence of succeeding populations in the same geography is 200,000 years old and continues up to our day.

Although humanity reached the American continents later than Europe or Asia, the various Indian populations also systematically replaced each other by the usual means of slaughter, displacement, or subjugation. The Sioux and the Crow chased and subdued each other. When Hernan Cortes conquered the Aztecs at Tenochtitlan, which is now the site of Mexico City, many neighboring tribes of different ethnicities that had suffered from Aztec domination, eagerly joined his efforts to expel and destroy their overlords.

The Serbs, Croats, and Bosniaks demonstrate a different variation in the failures of the racial classification scheme. Each of these three is identified and coded differently in the Census racial classification and each rolls up to a different subtotal. Yet, before the ninth century, all three were one people living in the same general area and totally indistinguishable. They still speak the same language, and even native speakers cannot usually distinguish the “native” tongue of their interlocutor. 

After the ninth century, as a result of geography, one group converted to Orthodox Christianity and became known as Serbs. A second group adopted Roman Catholic Christianity and designated themselves Croats. Later, in the 14th century, the Ottoman Empire conquered portions of both groups. That slice of the population converted to Islam and became known as Bosniaks. The three differ from each other only by their religious heritages. Because religious scriptures were the first widely available written documents, the secondary effect was that in written communication the Serbs used the Cyrillic alphabet, the Coats the Roman, and the Bosniaks a mixture of both, abandoning the Arabic script used during their occupation by the Ottoman Empire. 

Since the only distinction is a religious one, asking someone who identifies as one of these ethnicities is equivalent to asking them about their religious affiliation. That directly violates 13 US Code § 221(c): “Notwithstanding any other provision of this title, no person shall be compelled to disclose information relative to his religious beliefs or to membership in a religious body.”[9] This is one example known to the author, but there are likely similar cases in other geographies inhabited by hundreds of ethnicities. This feature alone should be enough to eliminate all racial and ethnic questions from government data collection.

The foregoing descriptions are illustrative only. There are undoubtedly hundreds more that make the same point, namely that there is virtually no place on earth where we know who the original inhabitants were, much less are able to trace any living person to one of them.[10] As part of our rich civic society, people celebrate all sorts of affiliations that they believe represent their ethnic and racial heritages. We should acknowledge and enjoy the richness that they bring, but designating just what categories should be observed is not within the government’s proper domain. Government by definition invokes violent force to gain compliance. 

Our individual understandings of who we are and where we come from are just that—individual and voluntary. When ethnic identities are defined and manipulated by government, they become instruments for oppression. The French learned that lesson from their Vichy genocide, and for most purposes they outlaw classifying official statistics by race or ethnicity. We should have learned the same lesson. It is now time to apply that learning.
 

Scott Lincicome

On February 7, President Donald Trump announced his intention to unveil a “reciprocal tariff” system sometime next week, so that the United States is “treated evenly with other countries.” Trump didn’t provide details of how such a system would work. But it appears he wants US tariffs on countries’ imports to match their tariffs on US exports: “They charge us. We charge them. It’s the same thing.”

Such a system raises a host of legal, economic, and practical concerns, but, being a Friday afternoon and all, I just have two basic questions for now: 

First, does this mean that the US government will reduce tariffs on imports from the many countries that have lower tariffs than we have?  As Figures 1 and 2 show for trade in manufactured goods and other non-agricultural goods, the United States does not have the lowest tariffs in the world and instead ranks somewhere in the middle among wealthy, industrialized countries. (Tellingly, only really poor countries have very high tariffs, but that’s a discussion for another time.) Should Trump’s system be based on average tariff rates, then “true” reciprocity would require US tariff rate reductions on goods from dozens of countries. 

As seen in Table 1, moreover, a country-to-county comparison would raise similar questions. There, we see that tariffs applied on US non-agricultural exports by the United States’ top export markets are about the same—sometimes a little higher, but also sometimes a little lower (see bold/​italic numbers)—as those applied by the United States on the same goods from the same places.

By this metric, a new reciprocal system means reduced US tariffs on manufactured goods from Europe, Mexico, Canada, or the UK, where, depending on how you measure the average, tariffs are actually lower than what the United States applies.

Second, will there be reciprocity for the other protectionism?  As I wrote a few months ago:

…Washington uses many “non-tariff” measures to impede foreign competition. This includes subsidies, quotas, “Buy American” restrictions, the Jones Act (drink!), and regulatory protectionism like the FDA’s blockade on baby formula. We’re also one of the biggest users of “trade remedy” measures (anti-dumping, especially) and today apply more than 700 special duties on mainly manufactured goods like steel and chemicals.

Indeed, according to the independent Global Trade Alert, which monitors nations’ trade liberalization and protectionist policies, the United States has had the most “harmful” trade interventions—tariffs, non-tariff barriers, subsidies, etc.—of any nation since late 2008. 

One can reasonably quibble with the organization’s methodology, but that wouldn’t change the basic point: many foreign nations do have protectionist barriers against American goods, services, and investment, but the United States is no saint here either. Instead, all countries’ trade policies have their political sacred cows (sometimes literally targeting cows!), and some countries are surely better than others. But the idea that the United States must adopt “reciprocity”—and in the process become a tariff-wielding protectionist outlier—because of some grave historical trade injustice is just detached from reality.

Jack Solowey

In 2023, some creative tinkering produced what you might call primitive “AI agents.” They were programs (with names like “BabyAGI”) that used LLMs to help users accomplish multistep tasks by breaking out and sequencing subtasks. The tasks were limited (mostly text-based) and performance was halting. But the app designs were clever, and the implications were vast.

I considered how a world of ubiquitous and capable AI agents—many of them open source—would upend traditional financial regulation if they began providing services like investment advice. Decentralized adviser bots, I argued, would require a decentralized governance framework—specifically one based on the common law of agency.

While the long-term governance issues never went away, AI agent hype died down for a bit. But following another round of AI progress, agentic bots are back. And a confluence of advancements along three dimensions—(1) reasoning models, (2) the open-source frontier, and (3) digital asset integrations—has made the need for, and the possibility of, decentralized governance for AI agents all the more compelling.

Agentic AI x Open-Source x Crypto

The integration of open-source reasoning models with digital asset tools heralds a future where increasingly autonomous software programs can provide retail financial services, from advising to trading. To understand how this might unfold, it’s important to take stock of the contributing factors.

First, agents are becoming mainstream. At the end of January, OpenAI launched Operator—an AI agent for performing web-based transactions (e.g., book a tour or buy groceries)—as a preview for Pro-tier users. And even before Operator launched, agentic AI was reemerging.

The LLMs debuting around Q4 2024 (like o1) specialized in so-called “chain-of-thought” reasoning, which lent itself to multistep tasks. While it’s a dramatic oversimplification of what’s known (not to mention what isn’t) about how these models were trained, reinforcement learning helped them gain their own expertise at the task decomposition and prioritization that apps like BabyAGI tried to achieve with after-market add-ons.

Second, advanced capabilities are becoming more accessible. Open-source reasoning models are now available, following the January debut of R1 from Chinese AI firm DeepSeek. As Miles Brundage writes, the lesson of R1 is that “AI capabilities that rival and ultimately exceed human intelligence are easier and cheaper to build than almost anyone can intuitively grasp, and this gets easier and cheaper every month.” Therefore, attempts to keep good enough AI genies in closed-source bottles will likely fail.

Lastly, these genies are probably going to have crypto wallets. o1 successfully spun up a Bitcoin wallet with a near-perfect success rate. Chatbots can call functions initializing digital asset transfers from non-custodial wallets based on user conversations. Open-source crypto AI agent frameworks are being built to connect LLMs to crypto market data sources, wallets, and smart contracts. As Shaw Walters, the developer of one such framework (ElizaOS) puts it, “If LLMs are kind of the brain, I think that Eliza and frameworks like it are really the body.”

While it’s still early days, and many so-called “crypto AI agent” activity involves LLMs posting weird memes on X, there’s reason to think financially empowered AI agents will be significant economic actors. LLM-based agents that outperform more traditional algorithmic trading agents (“in terms of Cumulative Return and Sharpe Ratio”) already have been built with relatively primitive models, like GPT‑4 Turbo (and some clever memory architecture). As frontier models continue to improve in general reasoning ability, quantitative analysis (a previous shortcoming), and coding ability, we should be prepared for their trading, advising, and financial engineering skills to improve as well.

Autonomy and its Discontents

Defining agency. We should always be wary of overreading promising AI agent signals. The pace of progress (and disillusionment) is quick, the takes are hot, and the cloud of marketing puffery is thick.

Relatedly, one of the greatest challenges for making sense of AI agents’ trajectory is definitional. Distinct but related terms—like autonomy, agency, and automation—are thrown around with imprecision, and our implicit thresholds for things like autonomy vary.

To ground the conversation, it’s worth defining terms. By my lights, an AI agent is an application that can (1) influence a virtual or physical environment, (2) at the direction, and on behalf, of a user, (3) autonomously, while (4) adapting to a dynamic environment. (This definition is based on a few sources: an OpenAI governance paper, Stuart Russell and Peter Norvig’s AI textbook, Drew Hinkes, and Vitalik Buterin).

In this definition, autonomy refers to the agent’s independence from both ongoing human intervention and exclusive reliance on its initial programming. Notably, this definition therefore covers agents that are not yet fully autonomous. Think of a wealth management agent that could independently research and prepare a few bespoke investment strategies but requires a steer toward the best one for you.

Autonomy is a spectrum. Self-driving vehicles already are judged on a 5‑level autonomy spectrum. A Level 0 vehicle has no autonomy, and a Level 5 vehicle has full autonomy. Level 1 and Level 2 vehicles have increasing numbers of semi-autonomous features (e.g., lane centering and adaptive cruise control). Level 3 vehicles do not require constant driver supervision. Level 4 vehicles do not require any driver supervision under certain road conditions. And Level 5 vehicles do not require any driver supervision on essentially all road conditions.

To rationalize the AI agent discourse generally, we should adopt similar autonomy spectrums in other domains, like financial services. Where an AI agent falls on that spectrum should inform its governance framework.

The governance challenge. Increasingly autonomous software, and open-source versions in particular, challenge the traditional regulatory model. In that framework, regulators set licensing and operating requirements for service providers (like investment advisers and brokers, in the financial context) who are typically able to cease and modify their services upon regulators’ instructions. However, when software development, deployment, and operation lack singular regulatory touchpoints, that governance model begins to break down. The long-term choice before regulators is either to usher in a new framework or attempt to force the technological genie back in the bottle. Even if the latter were possible (it’s probably not), it would require intolerable degrees of surveillance and prohibition. Reform is in order.

Reform

The wisdom of the common law. Unlike the above-mentioned framework of licensure and ongoing compliance prescriptions, a common-law approach would allow unobstructed market entry while incentivizing adherence to reasonable standards of care.

Not only does this decentralized governance approach accommodate ubiquitous autonomous actors, but also its learning mechanism is well-suited to governing uncharted territory.

As Friedrich Hayek argued, the common law was not “invented” but discovered. It gradually accreted wisdom by selecting for the most adaptive doctrinal developments. The output of the common law judicial process, in Hayek’s words, is “the experience gained by the experimentation of generations,” which “embodies more knowledge than was possessed by anyone.”

Lest one think that Hayek committed “the naturalistic fallacy (that is to make the claim that, whatever evolves, is good),” he pointed to the beneficial outcomes of the common law regime, not merely the elegance of its learning algorithm. Specifically, the common law helped to achieve a spontaneous order that was conducive to mutually beneficial relationships and resisted tyrannical abuses of power. We should not only welcome but strive for a similar outcome in a society of autonomous humans and AI agents alike.

Society’s reinforcement learning. A common law governance framework for autonomous AI makes sense for two interrelated reasons: one, it would assist with rule discovery, and two, its enforcement mechanism parallels a key way models underlying AI agents learn.

An iterative process for discovering standards would help address the knowledge problem presented by novel AI agents. In the words of Russell and Norvig, “It is impossible to anticipate all the ways in which a machine pursuing a fixed objective might misbehave.” Therefore, we will need a means of translating the lessons of experience into future rules.

Second, the parallels between common law remedies and reinforcement learning make the common law the most conceptually apt framework for AI agent governance.

While modern techniques are highly nuanced and always being refined, Russell and Norvig explain that in reinforcement learning at its most basic:

“[T]he agent learns from a series of reinforcements: rewards and punishments. For example, at the end of a chess game the agent is told that it has won (a reward) or lost (a punishment). It is up to the agent to decide which of the actions prior to the reinforcement were most responsible for it, and to alter its actions to aim towards more rewards in the future.”

The common law could be thought of as society-wide reinforcement learning. Reasonable behavior is rewarded with society’s blessing; unreasonable behavior is punished with financial and other penalties. Unlike in chess, the underlying rules that make human relations, in all their messiness, harmonious are not necessarily known in advance. As alluded to above, they are learned by hard-won experience, codified through institutions like the common law.

Fortunately for AI agents, they can learn directly from simulations (like self-play in chess), as well as experience implicit in datasets before entering the wild. At present, their deficiency is in real-time, real-world learning.

Overcoming that technical challenge is far beyond my capabilities. But regardless of whether and how it’s overcome, AI agents’ behavior in the real world also will need to be evaluated for adherence to social and legal conventions. Such conventions have been established over centuries in the common law of agency. 

Lessons of agency law

The current body of agency law is neither perfect in itself nor perfectly adapted to AI agents. It evolved to accommodate the crooked timber of human nature. And while the fact that LLMs have been trained on large swathes of recorded human thought suggests LLM nature may resemble human nature in important ways, there likely will be important differences.

In LLM development, pretraining is the process by which a model acquires basic language ability before being further finetuned for more sophisticated outputs. The current body of agency law could be analogized to a pretraining foundation out of which the common law of AI agency further evolves.

The current common law of agency is a useful foundation because it surfaced the questions raised by millions of principal-agent fact patterns (both the typical and the edge case). While agency law’s answers point to key principles for achieving socially efficient outcomes, the questions themselves are just as important. What follows is a taste of those questions, how the answers have tended to shake out, and their relevance to AI agents.

Agent capacity. The first question is whether an entity has the capacity to be an agent. There’s reason to think AI agents could one day. An agent need only be a legal person, not a natural person, and there’s historic precedent for assigning key features of legal personhood to machines. Moreover, even a human agent need not have all the rights of a legal adult before being considered an agent. A minor who lacks capacity to enter a binding contract may nonetheless do so on behalf of a competent adult principal. This has critical implications for identifying the technical threshold at which AI agents could be considered legal agents. In short, it probably will be before “Level 5” autonomy.

Who’s the principal? A key consideration for AI agents is who should be considered the principal when a user is employing an AI agent from a service provider. For instance, assuming the Operator could be considered a legal agent, would the user (i.e., the “special employer”) or OpenAI (i.e., the “general employer”) be responsible were the Operator to negligently injure a third party? Agency law addresses this question with the idea of the “lent employee” (or “borrowed servant” in the more antiquated argot).

For better or worse the test is flexible—and joint liability is possible. The primary question is which employer was in a better position to prevent the injury, which in turn comes down to who was in practical “control” of the agent. Answers are highly fact dependent. While the lack of a hard and fast rule can be frustrating—and is a downside of the common law approach—it also reflects and accommodates the messiness of reality.

Legal alignment. One of agency law’s greatest discoveries is the set of duties agents owe their principals. It defines, in a word, alignment. Aligned agents act: in good faith; within the bounds of their authority; according to all of the principal’s lawful instructions; and “with the care, competence, and diligence normally exercised by agents in similar circumstances.” In addition, aligned agents are loyal, not putting their own interests ahead of their principals’ interests. Before AI agent alignment can be achieved, it first must be put into words. Agency law refined those words over centuries of trial and error.

Liability. What many really want to know from agency law is when the principal is “vicariously liable” for harm caused by the agent. Traditional tests focus on the extent of the principal’s control, as well as the scope of the agent’s employment. Where a principal controls the goal, but not how the agent achieves it, the principal is less likely to be held liable. Similarly, the principal is less likely to face liability where the agent’s tort was so far outside the bounds of the job—for example, in terms of time and space or the nature of the activity—that it was serving the agent’s purposes, not the employer’s. While such tests are litigated with mixed results, at its best, agency law points to the efficient principle that vicarious liability only should arise where the employer was best positioned “to take cost-effective precautions.” Lest one think the principal only subsidizes risky agent behavior, when the agent breaches its fiduciary duties, the principal can recover from the agent. (More on how that could work in the AI agent context below.)

Implementing Reform

Getting to a world where AI agents are governed by a specialized common law will take continued technological advancement and creative legal reform. Here’s one way it could happen.

The growing thicket of regulatory requirements touching likely AI agent activity—such as the Securities and Exchange Commission’s investment adviser regime and state-level AI consequential decision laws—presents an obstacle to AI agent deployment, but also a reform opportunity in disguise. AI agent developers could be afforded safe harbors from liability and obligations under such regimes where they make their agents amenable to common-law remedies.

For example, developers could be incentivized to construct agents that accommodate legal injunctions, prioritizing instructions from duly constituted legal bodies notwithstanding a principal’s instructions. Similarly, the agents could be equipped with digital wallets capable of paying damages or indemnification where they’re adjudicated at fault for torts or in breach of fiduciary duties.

Where might the money come from when the AI agent is directly liable to a third party or a principal? The more autonomous an AI agent is, the more likely it is to possess funds. While principals and others often will be the beneficial owners of those funds, consider the following scenario. A decentralized group of open-source developers contribute code to a trading agent, which proceeds to operate with a high degree of autonomy. A percentage of the agent’s trading return runs to a foundation that assists with routine updates to the codebase. The agent is also available to retail traders. Some traders may pay the agent a fee for its services, others may allow it to retain a percentage of profits. Either way, where the agent’s revenue exceeds its obligations to beneficial owners and the foundation, it will have its own money. 

In addition, the developer safe harbors described above could require agents’ terms of use to subject the agents and their principals to arbitration in special-purpose alternative dispute resolution (ADR) fora, wherein the common law of AI agency can evolve. The fora might even be run by arbitrator agents trained in the common law of agency. Because litigation is an adversarial process, the fora can borrow from training concepts like self-play. So, in addition to learning through adjudication of actual controversies involving AI agents, AI common law also could advance through simulation—speedrunning centuries of trial and error. Recording real-world AI agent transactions on public distributed ledgers could provide useful data for both running simulations and understanding long-term outcomes in the wild.

The business end of AI ADR would be adjudicated rewards and punishments. Bot-specific cryptocurrencies could help allocate these rewards, augmenting whatever disincentive takes hold when AI agents fork over damages in a human medium of exchange. While it remains to be seen whether such external rewards and punishments could modify an individual AI agent’s behavior, at the very least, a public ledger of adjudicated AI agent rewards and punishments would seem necessary for the broader ecosystem to understand and update how agents perform in the wild. Recording that knowledge publicly could inform consumers and AI agents alike about which agents to trust and avoid. It would provide data to assist developers in pushing changes to their agent frameworks and adjusting relevant underlying models to correct past misalignment.

Conclusion

Human behavior is governed by a series of internal and external reward systems—from neurotransmitters to money. At its best, the external rewards system encourages pro-social and mutually beneficial behavior where mediated by a price system and the rule of law, including the institutions of common law courts. Ensuring pro-human and positive-sum behavior from AI agents will require further discovery and dissemination of standards. The common law of agency and public ledgers of digital rewards are two key tools for learning lessons and reinforcing them. If we want to avoid the twin dangers of Hobbesian chaos and Orwellian tyranny in a world of dizzying progress toward autonomous AI agents, we should pick up these tools.

David J. Bier

During his inaugural address, President Trump promised to deport “millions and millions of criminal aliens.” I have a new essay exploring this premise. In it, I note:

There are not “millions and millions” of removable noncitizens who have criminal convictions.
Trump has rescinded requirements to focus on public safety threats.
Trump is mandating ICE officers arrest a certain number of people per day, even if the officers think the arrests are unjustified.
One report found that only half of those arrested on one day last week had a criminal conviction or pending criminal charges.
A new report not available at the time of my piece includes data showing that a majority of the 8,200 arrested do not have a criminal conviction of any kind.
Trump is redefining an alien with a “criminal record” to mean anyone here illegally, even as he takes away people’s legal status to make them here illegally.
Trump is diverting criminal law enforcement to work on arrests of low-level immigration offenders.

You can read the entire piece here.

Chris Edwards

President Donald Trump is making rapid decisions on a wide range of issues. His fingers are into everything. Some actions are misguided and divisive, as were many actions of his predecessor. Modern presidents can be so disruptive because the federal government has amassed so much power over so many activities.

If you are bitter about federal politics, you should consider the libertarian solution, which is to devolve most current federal activities to the states and the people. That would ease tensions by strengthening local democracy and allowing for beneficial diversity across the states.

With Trump, for example, I’m all for him working with Congress to shut down the Department of Education and freeing the states to pursue their own school reforms. But I’m not fond of his recent actions to redirect water flows through California’s complex reservoir and aqueduct systems.

California’s water is supplied by a huge infrastructure of both federal- and state-owned assets. The balancing of water interests—fish vs. farmers vs. cities—is a political hot button. That is why the balancing should take place in water markets and Sacramento, not Washington, DC.

Rather than having a water czar in the White House, I’ve argued that Congress should end federal subsidies, transfer federal water assets to the states, and support water markets to encourage efficient allocation and use.

It is true that Los Angeles bungled its water readiness during the recent wildfires. Trump pounced, issuing an order to “maximize” water deliveries and “override” state policies if needed. Water was released from two federal reservoirs, and Trump proclaimed, “the water is flowing in California … heading to farmers throughout the state and to Los Angeles.”

However, CNN reports:

The newly released water will not flow to Los Angeles, and it is being wasted by being released during the wet winter season.

“They were holding extra water in those reservoirs because of the risk that it would be a dry summer,” said Heather Cooley, director of research for California water policy organization the Pacific Institute. “This puts agriculture at risk of insufficient water during the summer months.”

… “This release is extremely concerning,” Cooley said. “It’s providing zero benefit and putting California farmers at risk of water supply constraints in the coming months.”

California Department of Water Resources director Karla Nemeth told reporters that there was little coordination between federal officials and the state and local water managers for the Army Corps releases at the Terminus Dam at Lake Kaweah and Schafer Dam at Lake Success.

“These reservoirs were federal reservoirs, and the state of California was not part of the decision making in this instance,” Nemeth said. “We traditionally have a high degree of coordination at the operational level, which really wasn’t a part of this decision.”

… Los Angeles’ water sources are completely separated from the water system that Lake Kaweah and Lake Success supply. That water system flows into the agriculture-heavy Central Valley — where large farms grow nuts, citrus and grasses for animal feed, among other crops.

Newsweek had a similar take on Trump’s action:

“These kinds of shenanigans, they hurt smaller farmers,” Dezaraye Bagalayos, a local water activist, told the Los Angeles Times. Small growers have already been struggling, and the release of water from the dams means they will have less when they need it, Bagalayos said.

“The last thing in the world California water management needs is somebody like Trump calling shots when he doesn’t know how anything works,” Bagalayos said. “It’s making an already hard situation very, very difficult. We don’t have a lot of wiggle room in the state of California to be messing around with our water supply like this.”

… Climate and hydrology expert Peter Gleick told CBS: “There is absolutely no connection between this water and the water needed for firefighting in L.A. There’s no physical connection. There’s no way to move the water from where it is to the Los Angeles basin.”

Trump’s intervention in Western waters is somewhat like President Biden’s intervention in Western lands last year. Biden imposed a plan to smother 700,000 acres in the West with solar panels, which would have ruined landscapes and threatened plants and wildlife. Western leaders have long chafed at such heavy-handed federal actions on the lands within their states.

The country would be better off without Republican or Democratic presidents micromanaging resources and programs within the states. Congress should transfer most federal lands and water infrastructure to the states and private sector. And it should eliminate federal programs, such as education programs, that duplicate state-local activities. Decisions over resources and programs should be made as close to the people as possible.

Polls show that Americans have a better view of their state and local governments than the federal government, and they prefer state rather than federal control over education, housing, welfare, and other activities. Moving federal assets and programs to the states would go a long way to easing political tensions and restricting rash presidential actions contrary to local needs and preferences. 

See this study for a discussion of federal water policies, and this study for a discussion on reviving federalism.

Colleen Hroncich

With decades of experience helping kids with special needs—both professionally and as a mom—it’s not surprising that Lisa Mitchell created Everything Under The Rainbow, a hybrid homeschooling program for neurodiverse children. Lisa’s two children, now adults, are both neurodiverse and learned very differently from each other. “My daughter is the one that, quite frankly, she’s the one that thought of it. She said, ‘I need this,’” Lisa recalls. “Several years later is when I launched it with $200 to my name and I said, ‘Okay, God, this is what I want.’”

EUTR wasn’t inspired by pandemic school closures, but Lisa happened to be ready to go in 2020. She found a place where they could meet in person at the pastorium home of an old church, and she had 21 kids in the first month. There are currently 15 students ranging from six to 14 years old. Lisa says two of them are neurotypical while the rest are neurodiverse. She notes that, while many of the students have various challenges, they are often quite brilliant in some areas. For example, she has a six-year-old student who is learning to read but can already multiply fractions. 

To ensure everyone gets the individualized attention they need, EUTR has a six-to-one staff ratio. The faculty has a diverse background with a lot of experience in special education, behavioral support, and trauma.

In keeping with Lisa’s goal of meeting the individual needs of students, she offers several enrollment options at EUTR. “At first, most people chose the two-day option, but now I’m finding that many are adding the third day,” she says. For next year, she’s already hearing that many families are planning three or four days.

The daily schedule at Everything Under The Rainbow includes several breaks, a long lunch and recess period, and project-based learning. Incorporating active, hands-on learning and breaks is particularly important for the neurodiverse students Lisa specializes in helping. They also follow a shorter school day compared to most conventional schools, meeting from 9:00 a.m. to 2:00 p.m.

While the academic component is very important, Lisa’s mission is much broader than academics. She wants to provide a supportive atmosphere that will allow children who may struggle in other environments to thrive. There is also an emphasis on strong values and morals, such as integrity, compassion, respect, and responsibility.

Her experience with her own children and with helping other families dealing with similar challenges has given Lisa a strong passion to help serve children who have unique learning challenges. “Everything Under The Rainbow is in honor of not just children that learn differently, but it’s in honor of the parents that are riding that ride,” she says.

Lisa doesn’t know where her journey will take her next. Georgia has a new education savings account program that could help more students access Everything Under The Rainbow. Unfortunately, only students enrolled in public school are eligible, so it won’t help her current students—even those whose families have a hard time affording tuition on top of the taxes they’re paying to support the public school system. But, in time, it could give students who aren’t thriving in their assigned schools the chance to benefit from EUTR’s individualized education. 

Adam N. Michel

Last month, we kicked off the Cato Institute’s Congressional Fellowship in Tax & Trade Policy ahead of a year that has already been busy for both policy areas. The first four discussion sessions focused on tax policy, covering all the critical information you need to know as Congress decides how to address the year-end expiration of the 2017 tax cuts.

With our bipartisan cohort of senior congressional staff, Chirs Edwards and I covered tax code basics, major features of the Tax Cuts and Jobs Act (TCJA), the history of radical tax reforms, and international tax. Each discussion is summarized in four blog posts for those who want to follow along and learn about the US tax code. This post compiles all four of those blogs.

I’ve also shared these posts on my Substack, Liberty Taxed, where I cover the twists and turns of this year’s tax debate, share relevant links, and explain the complicated details of the US tax code.

Week 1: Tax Code 101

The first week started with an overview of federal revenue sources and the distribution of who pays taxes—the rich pay the lion’s share of federal taxes. The session also covered distinctions between income and consumption tax bases, the number and types of tax expenditures, and the economic inefficiency of high tax rates, which underlies the well-known Laffer Curve.

Read the full outline here: Tax Code 101.

Week 2: Everything You Need to Know About the TCJA

At the end of 2017, Republicans enacted the TCJA, making significant changes to nearly every major portion of the US tax code. Most of the 2017 reforms are set to expire at the end of this year. If no action is taken, taxes will automatically increase, affecting Americans at all income levels.

I gave our fellows an overview of the TCJA’s individual tax cuts and then reviewed the business changes and their economic benefits. The session wrapped up with a discussion of the bill’s fiscal consequences. The written outline ends with a summary of the biggest changes made in the TCJA and links to additional readings.

Read the full outline here: Everything You Need to Know About the TCJA.

Week 3: The Quest for Radical Tax Reforms

This session started with a brief history of major tax reform efforts, an overview of flat-rate consumption tax proposals from the past half-century, and the Estonian tax system as a real-world model of a simple, flat income tax. The discussion ended with a bit of self-promotion. The Cato Tax Plan offers lawmakers and their staff a detailed list of specific reforms to move the federal income tax to a flat tax with historically low tax rates on wages, investment, and business income.

Read the full outline here: The Quest for Radical Tax Reforms.

Week 4: An International Tax Primer

In our final tax discussion, we explored the complexity of the international tax system. I explained the competing goals of multinational taxes, outlined the three theoretical models of cross-border business taxes, examined how real-world rules align with these theories, and reviewed the extent of profit shifting, which is small and declining.

We concluded with an overview of the Organisation for Economic Co-operation and Development’s (OECD’s) two-pillar framework, introduced in 2020 as an attempt to curb profit shifting by raising taxes on American businesses and redistributing taxing rights away from countries like the US. President Trump distanced the US from the agreement early in his term, and his administration’s opposition is likely to stall its progress.

Read the full outline here: An International Tax Primer.

On to a Busy Year

This discussion series was an excellent way to kick off what promises to be a significant year for tax policy. The fellows asked insightful questions, demonstrating a strong understanding and interest in this important policy area. Bipartisan dialogues like these are essential for building a shared understanding and advancing better policy outcomes.

To stay updated as this year’s tax debate unfolds, subscribe to my Substack for tax analysis delivered directly to your inbox.

Congressional staff interested in participating in future Cato fellowship programming should contact GovAffairs@​cato.​org for more information.

Rachel Johnson contributed to this post. 

Gene Healy

Two weeks into the first Donald Trump presidency, Washington was freaking out about presidential style: “alternative facts” about Inaugural crowd size; scandalizing the CIA(!); using the National Prayer Breakfast to roast Arnold Schwarzenegger for lousy ratings as host of Celebrity Apprentice.

Eight years later, the Beltway meltdown is largely about policy substance. Little wonder: Where Trump 1.0 issued only a single first-day executive order, Trump 2.0 came out of the gate with 26—and the pace has barely slackened since.

President Trump’s flood-the-zone approach has led to some breathless hyperbole from legacy media: “madness”; “unprecedented times”; a “terrifying flurry”! C’mon man: Given the death-metal pace of executive orders at the start of Joe Biden’s presidency, Trump’s flurry is hardly “unprecedented.” As for “terrifying,” I suppose it depends on what scares you. For the political press and official Washington, whether “energy in the executive” is cause for celebration or panic turns entirely on who’s getting zapped.

Libertarians should take a more principled view. That’s my aim here: first, the bad—Trump’s most dangerous power grabs. Then, the bright spots—orders that advance liberty without shredding the Constitution. Finally, I’ll take a broader look at the dangers of our accelerating slide into pen-and-phone governance. 

Bringing Negative Energy

Trump’s executive-order onslaught offers plenty for libertarians to loathe: two bogus national emergency declarations, the militarization of border-control and drug interdiction policy, and the weaponization of the International Emergency Economic Powers Act in service of “the dumbest trade war in history.”

At JustSecurity, Adam Cox and Trevor Morrison flag three Trump directives that rest on a George W. Bush–style “constitutional prerogative to ignore, disregard, or even openly violate federal laws.”

the order purporting to rewrite the Fourteenth Amendment by eliminating birthright citizenship;
the executive order (EO) suspending the TikTok ban without a shred of statutory cover (a bridge too far even for John Yoo, author of the Bush torture memos); and
the “Invasion” Proclamation, which, per Cox and Morrison, conjures up an extralegal prerogative to suspend any law Congress has passed affording noncitizens the right to remain in the United States.

That’s all pretty bad, though it’s an open question how much of it will survive first contact with the federal courts (so far, the birthright citizenship order looks only slightly more viable than Biden’s 11th-hour attempt to graft the Equal Rights Amendment onto the Constitution). 

Dismantling the State Church of Woke and Other Policy Wins

The good news is, among the orders that are legally sound, there are some clear policy wins for supporters of individual liberty, limited government, and free markets. Chief among them are the directives (1) eliminating affirmative action in the federal government and for federal contractors (EO 14173); and (2) ripping out federal diversity, equity, and inclusion programs root and branch (EO 14151).

Those directives go a long way toward making good on Trump’s pledge, in his second inaugural address, to “end the government policy of trying to socially engineer race and gender into every aspect of public and private life.” Last I checked, that was at least as important a libertarian policy priority as freeing Ross Ulbricht. So why the long faces?

Any post–Lyndon B. Johnson Republican president could have ended affirmative action; except for Ronald Reagan, none of them even tried: too divisive! Wouldn’t be prudent! Credit where due: Trump actually did it. Some norms deserve to be smashed.

Of course, I’m not crazy about Section 4 of the “Ending Illegal Discrimination” directive (EO 14173), which orders each federal agency to find up to nine potential civil rights cases against private-sector entities. But they’re not going to have to dig very deep to fill that quota, given the extent to which many American universities and Fortune 500 companies have been flagrantly violating the law. It’s pretty clear, for example, that if AT&T adopted workplace training that insisted, “black people, you are the problem” and “racism is a uniquely black trait,” the Equal Employment Opportunity Commission would want to have a word. It’s not exactly state neutrality to only enforce the race-neutral Civil Rights Act on behalf of preferred groups. If you’d prefer that the feds get out of this business entirely, then consider whether Barry Goldwater had a point.

Other pen-and-phone policy wins include: 

EO 14168, which lays the groundwork for repealing the Biden Title IX rule governing bathroom access for practically every K–12 public school and college in America;
“Restoring Freedom of Speech and Ending Federal Censorship,” (EO 14149) forbids federal officials from “engag[ing] in or facilitat[ing] any conduct that would unconstitutionally abridge the free speech of any American citizen.” (A good first step, but as Andrew M. Grossman and Kristin A. Shapiro argue a 2023 Cato study, “The way to address censorship by proxy [is to] require government officials to disclose it”; such a requirement would actually give the order teeth); and
EO 14154, “Unleashing American Energy,” which announces the administration’s intention to “eliminate the [Biden] ‘electric vehicle (EV) mandate’” and roll back water-flow restrictions on showerheads and toilets.

Is This Any Way to Run a Country?

Taking a longer view, it is, of course, insane that 21st-century presidents get to decide what kind of car you can buy, who gets to use the girls’ locker room, whether you can smoke menthols, and how much water flows through your showerhead and commode.

It shouldn’t be this way; in living memory, it wasn’t. Thirty years ago, first-week stroke-of-the-pen policy changes were pretty small beer. When Bill Clinton took office in 1993, for example, he reversed two Reagan-Bush EOs: the Mexico City Policy on foreign aid and abortion and another ordering federal contractors to post notice of their employees’ Beck rights. Today, fundamental questions of governance turn on which team, red or blue, manages to seize the presidency.

If we’re stuck with this crazy system, then I prefer greater automotive choice, fewer bureaucratically enforced ideological manias, and, of course, luxurious, free-flowing showers and hearty flushes. But at some point, you have to ask yourself, is this any way to run a country?

In Federalist 70, Alexander Hamilton insisted that “energy in the executive” would foster “steady administration of the laws.” But living with the turbocharged modern presidency means whipsawing between extremes as the law changes radically whenever the office changes hands. That system isn’t just stupid, it’s dangerous: By raising the stakes of the transfer of power, it risks making every presidential election a “Flight 93 election.” 

There’s no shortage of smart legislative proposals for reining in presidential power. If we don’t want the president to be able to unlock new statutory powers by saying the magic words “national emergency,” Congress could amend the National Emergencies Act so those powers quickly expire unless Congress votes to approve the national emergency declaration.

Sen. Rand Paul’s REPUBLIC Act aimed to do just that—while also blocking the use of the International Emergency Economic Powers Act as a trade-war weapon. Other bills from Senator Paul and Sen. Mike Lee would require congressional approval for presidentially imposed trade restrictions. But while the REPUBLIC Act cleared committee last September, none of these proposals even made it to the floor—let alone the president’s desk.

The political problem is that, barring a veto-proof supermajority, these framework statutes require the president’s signature. And there’s no way President Trump is going to go along with statutory reforms that clip his wings. He likes declaring phony national emergencies and launching trade wars from his couch.

Another key reform, the REINS Act, may have a better shot at getting this president’s signature. The REINS Act (for Regulations from the Executive in Need of Scrutiny) would require “major rules” (those with an economic impact scored at $100 million a year or more) to get affirmative approval from Congress. Versions of the act have passed the House several times, including in the last Congress.

Some of Trump’s close advisers support the REINS Act, and it could be pitched to him as a way to frustrate future presidents who want to restart regulatory initiatives, like the Biden EV mandate, that he’s undone. It would also, by reining in the “imperial administrative presidency,” foster steadier administration of the laws.

As a recent survey shows, “Americans think a president’s power should be checked—unless their side wins.” But even if you’re currently enjoying the sowing, you’re going to get the reaping when power inevitably changes hands. In our partisan myopia, we’ve allowed the presidency to become powerful enough to tear the country apart. However daunting the current political obstacles to executive-power reform, there’s never been a greater need to rein in the president’s emergency powers, war powers, authorities over trade, and the ability to make law with the stroke of a pen.

David Inserra

Last week, Salwan Momika, an Iraqi immigrant to Sweden was murdered because he burned Qurans to show his opposition to Islam. His murderers remain at large, though the Swedish government said that the killing could be connected to a foreign power, similar to how author Salman Rushdie has been frequently targeted by Iran for his 1988 novel The Satanic Verses and was almost killed in 2022.

And yet, liberal, modern Sweden and the likely fundamentalist murderer or foreign actor were aligned in their belief that Mr. Momika and his expression were a problem and violence was needed to stop him. He was murdered shortly before he was set to be convicted by the Swedish state for his Quran-burning “hate crimes.” Mr. Momika’s fellow activist Salwan Najem was convicted days after Momika’s murder for his burning of Qurans and hateful statements, with the Swedish courts saying their actions “by a wide margin” had gone beyond acceptable religious criticism. 

Sweden got rid of its laws against blasphemy in the 20th century. But significant unrest in Scandinavia over the past few years has led to governments rolling back their protections for free speech to criticize religions, even with book burnings. Last year, Denmark fully capitulated, reestablishing formal blasphemy laws on its books. Sweden is instead expanding the use of its other hate speech laws to effectively act as blasphemy laws that prevent criticism of religious ideas, books, or other materials. 

While Sweden’s punishments are certainly less severe than other blasphemy laws around the world, they are still using the force of the state to compel and punish speech that some find offensive.

And it’s not alone. Facing various incidents of unrest of its own, the UK has increasingly resorted to censorship. The UK is prosecuting a man, Martin Frost, for hate speech after burning the Quran in solidarity with the murdered Momika after his daughter was killed in the Israeli-Palestinian conflict. Despite England having repealed its blasphemy laws in 2008, the judge found that “The Koran is a sacred book to Muslims, and treating it as you did is going to cause extreme distress. This is a tolerant country, but we just do not tolerate this behaviour.”

While burning Qurans may elicit thoughts of Nazi book burnings and other prejudicial views, these actions are ultimately peaceful protests by individuals burning their own property to criticize a religion with which they vehemently disagree. It’s no different than desecrating a bible, burning an American flag, or burning your Harry Potter books because of JK Rowling’s views of sex and gender or its sinful promotion of magic and demonic influence. Whatever their motivation, these acts may be offensive, but liberal societies do not criminalize offense.

Nearly everything can be offensive to someone. Fervent followers of various religions may find attacks on their beliefs, symbols, or holy texts deeply offensive. Atheists may find all religious influence offensive. LGBTQ groups may find traditional sexual mores deeply upsetting. Conservative and traditional religions may find LGBTQ sexual views corrupting and evil. One nationality may find the mannerisms or behavior of another nationality to be disrespectful, and vice versa. Men and women may fault each other for objectionable or indecent speech. Comedians may offend nearly everyone. Whether they are microaggressions, jokes, or purposeful burning of a religious text, offensive speech can also take any number of forms because it ultimately is in the eye of the offended. 

And so criminalizing speech because another group finds it offensive is to give into the most intolerant and illiberal parts of society. If applied broadly, laws against blasphemy and offense will leave us in a totalitarian state without any humor, art, or public discussion. But more commonly, hate and offensive speech laws are inevitably used selectively to silence viewpoints. While banning anti-Islamic speech is currently in vogue in these countries, nothing stops a change in government from turning these laws against Muslims. History is full of examples in which laws against offensive speech are used to silence minorities and the oppressed. 

Not only do such laws backfire against the groups they are meant to protect, but they fundamentally undermine tolerance across society. Mr. Momika’s prosecution and now murder inspired Mr. Frost in the UK, whose arrest will inspire others. And if peaceful speech isn’t allowed, more people will unfortunately turn to increasingly radical ideas, unrest, and violence as they have no lawful way to express their non-violent ideas. Indeed, research indicates that free speech is essential to reducing social conflict as it gives the dissatisfied in society a way to make their voice heard without violence. 

Even for those who don’t go on to engage in criminality, blasphemy and hate speech laws cannot force people to respect ideas they do not respect. If anything, these prosecutions are likely to drive an even deeper wedge between Muslims and the rest of society, which will wonder why Muslims need special legal protection for their ideas.

The state of free expression in Europe and around the world is growing more and more concerning. And the ideas being popularized and enacted likely are having an impact on how Americans view expression. So it is important that liberty lovers call attention to this worldwide backslide on free expression.