Honesty

October 19, 2025

By Stephen Stofka

ChatGPT was released just three years ago and AI technology has been widely adopted by the public. Some high school students use AI programs to write their college admissions essay. These essays are easy to identify and may affect a student’s chance of acceptance (Source). The AI detection tool Turnitin, used by many colleges, has identified some AI text in millions of students’ papers (Source). What’s the concern? Students are judged on their efforts, not the efforts of OpenAI, the company that developed and trained ChatGPT. In this context, misrepresentation is dishonest. The issue is not AI, but honesty. This week I want to explore the axis of honesty and dishonesty. My second axis will be personhood, whether an entity is regarded as a person or not.

Dishonesty is a master of disguises. Sometimes we are dishonest to protect our privacy. A company asks for a phone number and we give them an old one or make up a number. We are dishonest to avoid embarrassment. We use dishonesty as a social grease, smoothing out uncomfortable interactions. We call them white lies. Parents struggle to explain to their young children the difference between a bad lie and a white lie. An older child, perhaps eleven or twelve, will argue with us over our criteria.

Politicians use dishonesty as a political grease to ease out of an uncomfortable or embarrassing set of circumstances. Corporations are dishonest for the same reason. They may be protecting trade secrets or disguising their real intentions from competitors. A company may announce the launch of a new product that they know is not ready to ship. They hope to attract more investment or dissuade competitors from entering that particular market. In 1985, a company called Ovation introduced a new suite of office software with a fake demonstration of its capabilities, even though the software was still in early development. The company hoped the demonstration would attract investment (Source). Esther Dyson coined the term “vaporware” to describe such products.

Some of us don’t respect corporations in general, but most companies have to operate with some degree of honesty. They have to supply a consistent product or service to retain current customers and attract new customers. They must meet their customers somewhere in the distance between the company’s needs and their customers’ needs. A corporation may behave like a ruthless profit machine at times but it not a robotic machine like an LLM.

Corporations are steered by people who act with intention. An AI makes connections between patterns of what appear to it as binary digits. We recognize the mistakes that AI programs make as hallucinations because we regard an LLM like ChatGPT as a machine, a tool. It has no intention, no driving force other than electricity. Its resources, the electricity, the chips and the digital storage, are given to it. If an AI appears to act with intention, it is because people imbue it with intention. My cordless drill just turns. It has no intention to make a ¼” hole in a wood panel. I am the one who transforms the turning of the drill into a hole.

Neither an AI nor a corporation are biological entities, yet the Supreme Court has ruled that corporations are persons. The issue first arose shortly after ratification of the 14th Amendment, which granted equal protection under the law to “persons.” In an 1882 Supreme Court case, San Mateo County v. Southern Pacific Rail Road,  Roscoe Conklin represented the defendant in the case, Southern Pacific Railroad. Conklin was a former Senator and the last living member of a Congressional Committee that had drafted the 14th Amendment. He informed the Court that the drafting committee had changed the word citizen to person to include corporations under the Equal Protection Clause of the amendment (Source).

In her book These Truths: A History of the United States, historian Jill Lepore (2018, p. 338) writes that Conklin was probably lying. Other evidence during the drafting of the amendment doesn’t support Conklin’s claim. Adam Winkler (2018), the author of We the Corporations, recounts the history of the corporate rights movement and concurs with Lepore. However, the Court made reference to that claim in another case four years later. In subsequent 20th century cases, the Court granted that corporations were artificial persons, but persons nevertheless for the purposes of the Equal Protection Clause (Source). That precedent was the basis for the Court’s decision in Citizens United giving corporations free speech rights. The court ignored any evidence that Conklin was lying because the conservative justices on the Roberts’ court fancy themselves to be very knowledgeable in both history and the law.

Conklin’s deception illustrates the fact that dishonesty can be an effective strategy to achieve one’s goals. In Conklin’s case, the likelihood of a favorable ruling from the court led to a settlement. Dishonesty as an effective strategy is only possible if we can anticipate the response to our dishonesty. We have to put ourselves in the mind of our audience, a difficult task for a young child, who may be befuddled when challenged over a ridiculous lie. “You didn’t do your vocabulary homework because it got wet in the rain. How is that possible? It hasn’t rained today.”

Animals may exhibit displays that misrepresent their size, or the direction they are facing but we don’t regard animals as dishonest. Many animals, particularly males, resolve disputes with aggressive displays that threaten violence to their opponent. Often, the display itself resolves the problem. One animal backs down out of concern for their own self-preservation. A display rather than an actual fight conserves energy.

Politicians lie so often that the public may be partially desensitized. But the police can lie as well? In a 1969 case Frazier v Cupp, the Supreme Court ruled that police can, in some circumstances, mislead a suspect to elicit a confession (Source). According to Standard 3-1.4 of the Criminal Justice Standards, a prosecuting attorney “should not make a statement of fact or law, or offer evidence, that the prosecutor does not reasonably believe to be true, to a court, lawyer, witness, or third party, except for lawfully authorized investigative purposes” (Source). Notice that law enforcement are bound by a negative. They “should not.” As the American Bar Association says, the word should is aspirational.

In court a lay or expert witness must take an oath to tell the whole truth. They are bound by a positive law that is not aspirational. Why the different standards? A prosecuting attorney is not supposed to give evidence. That is the duty of a witness. However, it can appear that there is a lower standard for law enforcement. This creates distrust of the police and the prosecuting attorneys that represent the government. Cynical public opinion might reason that the job of law enforcement is to get a high conviction rate in order to be an efficient use of taxpayer money. Their job is not to tell the truth. This raises another issue. Do our expectations of honesty vary with the role that people play?

I do not expect my friends to be honest in a Friday night poker game. I do expect them not to cheat. We make a distinction between honesty and cheating. In poker dishonesty is a guarding of private information, the value of the cards in my hand. The government sometimes plays a form of poker with the public. Accused of hiding information, a government official will claim that there are security concerns. The information may reveal incompetence, or poor decision making and the administration wants to protect its reputation. An opposing political party might use the information to sway public opinion in an upcoming election. Is the government cheating?

The Freedom of Information Act was passed in 1966 and amended in 1974 after the Watergate scandal. The act applies only to federal agencies. Each state has its own open records procedures. A media outlet first files a request for information from a U.S. government agency. If the agency refuses the request, the requester can file an appeal with the agency. If the agency denies the appeal, then the requester of the information can sue the agency in a U.S. District Court (Source). As in a poker game, the requester must expend both time and effort to access the information. Sometimes the government redacts a lot of the information contained within the released documents (Source).

There are also instances where government agencies are required to not release information  which would interfere with the pricing mechanisms of the stock and bond markets. The Federal Reserve has a blackout period of about ten days before an FOMC meeting to determine interest rates. Here is a link to this year’s calendar. During this period, Fed officials do not respond to policy questions from the media so as to not create expectations that might influence the market.

The Bureau of Labor Statistics safeguards the information in the monthly labor report before releasing it at a specific time, usually on the first Friday of each month. Some employees at the White House are advised of the data before its release so that they have comments prepared in advance (Source). In 2018, Trump moved markets when he tweeted a hint at the numbers in advance of the official release time (Source). In this case, a violation of the rules had the same effect as cheating. Trump has also been accused of fudging his golf handicap (Source). The President may regard both dishonesty and cheating as effective strategies to reach his goals, but I would not play poker of golf with someone who had that attitude.

When governments or corporations hide information, we may call it dishonesty, or a lack of transparency. Which is it? The label depends on the label maker. If we are hostile to the policies or actions of a particular government, we may label that dishonesty. Someone who is more neutral might label it a lack of transparency. Someone who favors that policy, party or government may speculate that there must be a good reason.

As it turns out, honesty is not easy to objectively identify in many everyday circumstances. I underline the word objectively. Even when we identify dishonesty, our reaction to it may be rather benign. We create judicial institutions and procedures to help unravel the subjective perceptions of honesty in a particular circumstance. In court we ask juries to evaluate the honesty and credibility of testimony. Unlike honesty, we have left the determination of personhood to a few justices on the Supreme Court and a footnote mention of corporate personhood in an 1886 case. Why do we treat these two concepts differently?

//////////////////////

Photo by Phạm Trần Hoàn Thịnh on Unsplash

Lepore, J. (2018). These truths: A history of the United States (First edition). W. W. Norton & Company.

Winkler, A. (2018). We the corporations: How American businesses won their civil rights. Liveright/Norton.

Claims to Truth

August 4, 2024

by Stephen Stofka

In this week’s letter I will clean up my notes on the judiciary before I start a series on investing next week. For several millennia, human societies have struggled to figure out the how and the who of lawmaking. How to construct a rule whose language is clear enough to be understood but flexible enough to apply to varying circumstances? Who will enforce the rules? How will they be enforced?

Let me start with a real story from my childhood. My younger brother and I were only one and a half years apart in age and were quite competitive. Tired of listening to my brother and I argue over the rules of various card games we played, our dad bought us a book called Hoyle’s Official Book of Games. This resolved some of our disputes, but still we argued over the interpretation of the rules. If a rule in Rummy requires a player to play a card once they have touched it, how to interpret the word “touch?” If a gust of wind threatens to toss a few cards from the deck, can a player reach out to prevent that without having to take that card? What if a player grazes the card with their elbow while reaching for a drink?

Former Supreme Court Justice Stephen Breyer (2024) wrote that many cases the court decides can hinge on the interpretation of one or two words in a law. In the Heller decision on the Second Amendment, the majority and dissenting justices wrote 80 pages of argument over the meaning of “bear arms” in the Second Amendment. He quotes the example of a railroad that required passengers to pay full fare for each animal they brought on board. When a biology teacher brought a number of snails on the train for a class, the conductor charged the teacher a fare for the snails. When the teacher complained that the rule was meant for pets, the conductor explained that the rule used the word “animals,” not “pets.” Snails are animals.

If you were a judge, how would you rule? Breyer’s story raises the question: what if the rule said, “domestic animals?” What does the term “domestic” specify regarding animals? An animal related to a person’s home or family. Snails can be kept at home so even this change in the rule could be interpreted to include snails. The most important symbol in mathematics is the equals sign. There are specific rules of valid operations on either side of an equals sign. Plain language has no equals sign. For centuries, French was the language of diplomacy. The language has fewer words, which conveniently left agreements between diplomats open to interpretation.

If the truth were an island, it would be stained with the blood of millions. Throughout history, kings, philosophers and countries have laid claim to and fought over the truth. In 17th century Britain, a civil war erupted between monarchists, those loyal to the king and the power of the king, and Parliamentarians, those loyal to the Parliament as representatives of the people’s will. Robert Filmer, a noted monarchist, argued that the right to property comes from God through the King. John Locke argued that the right to property comes directly from God. When God drove Adam and Eve out of the garden, He told the couple they would have to work the land for their survival. Therefore, those who cultivate the land add value to the land and “thereby makes it his property” (PDF link, p. 116). The state merely recognizes that title. In later centuries, the American colonists would use this reasoning to deny that Indians had any right to the land because they did not cultivate it.

Journalists often make claims on the truth. Matthew Yglesias wrote on his Substack “this is journalism, and we owe a duty of truth to our audience.” He referred to “journalists’ obligations of candor,” but truth and candor, or honesty, are different. In its 1964 decision New York Times v. Sullivan, the Supreme Court ruled that statements about public officials were not subject to claims of libel unless they were shown to be made with “actual malice” and a disregard for the truth. That is a standard of candor, not truth.

The American Bar Association states that lawyers “must be honest, but they don’t have to be truthful.” They distinguish between being honest and being truthful. “A defense lawyer has no obligation to actively present the truth.” Being honest requires only that “Counsel may not deliberately mislead the court.” A lawyer who serves their client’s interest by making a false representation to the court risks disbarment. Rudy Guiliani was recently disbarred for making deliberate misrepresentations to further Trump’s claims of a stolen election.

A witness, on the other hand, must vow to “tell the whole truth” to the court. Politicians take no such vow, nor do they pledge a vow of candor, to be honest with the public. They take an oath of office to “support and defend the Constitution against all enemies,” a vague statement open to wide interpretation. It is up to the public to judge their lies, and to assess the importance of those lies. Abraham Lincoln supposedly claimed that all the people could not be fooled all the time, but a politician does not have to meet that high bar. They only need to fool enough of their constituents to get re-elected.

Should Congress set minimum years of service as a judge for Supreme Court nominees? Chief Justice John Roberts, and Justices Kagan and Thomas had the fewest years on the bench before joining the Court. I will leave the details and links in the footnotes. I have many more notes after reading Breyer’s recent book, but I will close this for now. Next week, I hope to present some perspective on investing in plain language without equations.

///////////////////

Photo by John Cameron on Unsplash

Keywords: rules, honesty

Breyer, S. G. (2024). Reading the constitution: Why I chose pragmatism, not textualism. Simon & Schuster.

The link to the justices’ bios is here. Chief Justice John Roberts served as an Appeals Court judge for only 2 years before becoming Chief Justice. Justice Clarence Thomas – one year. Justice Samuel Alito – 16 years. Justice Elena Kagan had no prior judge experience. Justice Sonia Sotomayor – 16 years. Justice Neil Gorsuch – 10 years on an appeals court.. Justice Brett Kavanaugh – 12 years. Justice Amy Coney Barrett – 3 years. Justice Ketanji Brown Jackson – 9 years.

Honesty, Power and Crypto

July 2, 2023

by Stephen Stofka

This week’s letter is about lying and some types of lies, the key part they play in our society and the steps we take to uncover and counter lying. This week’s picture by David Clode on Unsplash is that of a butterfly wing, not an owl.

Like many animals, humans survive by signaling. A growl, the curl of a tail or a frown on a face are forms of signaling. Lying is a signaling tool that we use to get something we want. We may want protection from some threat so we lie. We may want approval from others so we lie. Lies come in several colors. Lies told for some social or public purpose are called white lies. In the 1969 Frazier v. Cupp decision, the Supreme Court ruled that the police could make false statements to gain a confession. These are known as blue lies. There are red lies told to hurt someone or their reputation. We tell green lies to gain some financial advantage.

People are so good at lying that public agencies and private companies spend billions per year to prevent fraudulent claims. As Jennifer Pahlka (2023) noted in her book Recoding America, policymakers will spend far more than a $1 to prevent paying out $1 of a fraudulent claim even if it means that legitimate claimants have to wait longer to receive their benefits. Policymakers are rarely as responsive to people’s needs as they were during the pandemic. In a rush to serve millions of people laid off during the pandemic, many fraudulent claims flooded state unemployment offices. According to an indictment filed in May, recently elected Congressman George Santos (R-NY) was one of those who made a fraudulent claim with New York State and received $24,000 while already employed.

Politicians like Santos attract news coverage but it is not clear that they are any more dishonest than the population in general. It is true that those who are uncomfortable with lying are cautioned not to run for office. Lying is an accepted tactic to confound the opposition, gain a policy foothold or some electoral advantage. Within a democracy, the electoral process is a competition of lies and boasts that political scientists call branding. There is no “truth in advertising” standard that politicians must adhere to when they run for office. It is the voters who must beware when they “buy” whatever ideological concoction a politician is selling. The voters are supposed to act as a giant sieve, straining out the fabrications, the incompetent and crackpots. It is not a perfect system.

The Bitcoin algorithm was designed to “crowd-source” property claims, spreading the verification process to the many nodes in a historical transaction chain. Yet the 15 year history of Bitcoin and other digital currencies has been punctuated with episodes of large scale fraud. According to a Justice Dept. investigation, in 2011, customers of Bitcoin exchange Mt. Gox learned that most of the bitcoin stored on the exchange had disappeared. Over a three year period, Russian hackers with unauthorized access to the exchange’s server had stolen most of the Bitcoin stored there. Dishonesty 1, digital security 0.

Forms of digital communication like email allowed scammers to send a lie for a fraction of a penny, far below the cost of bulk mailing. The Federal Trade Commission reported that consumers lost $8.8 billion to fraudsters in 2022. Money transfers, both legitimate and criminal, happen with the flip of some ones and zeroes. Digital currencies can be stolen at far less personal risk than holding up a physical bank so it is surprising that more crypto is not stolen each year. CNN reported an estimate that $3.8 billion worth of digital currency, most of it DeFi, not Bitcoin, was stolen in 2022. That’s just 0.45% of the $840 billion in the market cap of cryptocurrencies at the end of 2022, as tracked by Coin Gecko. The technology that underlies digital currencies could be adapted to verify other forms of transactions.

A century from now, we may put digital currencies in the same historical bucket with the worthless stock certificates of hundreds of railroads and mines issued in the 19th and early 20th century. History is littered with broken dreams destroyed by deceit. Dig down to the ideological foundations of digital currency, however, and there is an enduring idea that will outlast whatever the current form of digital currency trading and transfer. That idea is as old as the Constitution – checks and balances. Money is information and information is power. Unless that power is checked, it accrues into an autocratic regime or an economic monopoly. Digital currency represents a yearning for a check on the accumulation of economic and political power. That idea will not go out of fashion.

///////////////////

Show 20230702ButterflyWing.png

Photo by David Clode on Unsplash

Keywords: honesty, crypto, Bitcoin,

Pahlka, J. (2023). Recoding America: Why government is failing in the Digital age and how we can do better. Metropolitan Books, Henry Holt and Company.