Richard Stevens

Fellow, Walter Bradley Center on Natural and Artificial Intelligence

Archives

AI in the Courtroom: How to Program a Hot Mess

Could AI make competent judicial choices in the court?
Imagine we’re assigned to design the artificial intelligence (AI) software to carry out legal analysis of cases like a human judge. Our project is “CourtGPT,” a system that receives a factual and legal problem in a case where there are two opposing parties, analyzes how certain statutes and other legal principles apply to the facts, and delivers a decision in favor of one of the parties. CourtGPT will make “legal decisions,” not decide “jury questions of fact,” and thus will function like a judge (not juror). To write a computer program of any complexity, we start by describing the entire program’s operations in English (my native tongue). Pro tip: If you cannot describe how your program operates in human language, then you cannot write a program in computer

Love Thy Robot as Thyself

Academics worry about AI feelings, call for AI rights
Riffing on the popular fascination with AI (artificial intelligence) systems ChatGPT and Bing Chat, two authors in the Los Angeles Times recently declared: We are approaching an era of legitimate dispute about whether the most advanced AI systems have real desires and emotions and deserve substantial care and solicitude. The authors, Prof. Eric Schwitzgebel at UC Riverside, and Henry Shevlin, a senior researcher at the University of Cambridge, observed AI thinkers saying “large neural networks” might be “conscious,” the sophisticated chatbot LaMDA “might have real emotions,” and ordinary human users reportedly “falling in love” with chatbot Replika.  Reportedly, “some leading theorists contend that we already have the core technological ingredients for

ChatGPT: Beware the Self-Serving AI Editor

The chatbot "edits" by reworking your article to achieve its own goals, not necessarily yours
My article, Utopia’s Braniac (short title), reported results from experiments showing that for one, ChatGPT actually lies, and secondly, it gives results plainly biased to favor certain political figures over others. I next ran a follow-up experiment: asking ChatGPT to “edit and improve” the Utopia’s Brainiac manuscript before submitting it.  Close friends told me they’d used ChatGPT to improve their written work and said the process is easy. So, I tried it myself on February 6, 2023. I entered “Please edit and improve the following essay” and pasted my piece in full text (as ultimately published). In under a minute, ChatGPT delivered its edited and revised copy. What did it do? I. Deleted Whole Section That Gave Readers an Everyday Context Utopia’s

Utopia’s Brainiac? ChatGPT Gives Biased Views, Not Neutral Truth

Look at what happens when you try to get ChatGPT to offer unbiased responses about political figures
Do you trust your pocket calculator? Why?  Maybe you’re using the calculator app on your phone. Enter: 2 + 2. You get an answer: 4. But you knew that already. Now enter 111 x 111. Do you get 12,321? Is that the correct answer? Work it out with a pencil. That answer is correct. Try 1234 x 5678.  My calculator app returns 7,006,652. Correct? I’m not going to check it. I’m going to trust the calculator. And so it goes. The harder the problem, the more we trust the computer. That’s one reason why many people trumpet the powers of Artificial Intelligence (AI) systems. Those systems can give answers to problems we individuals couldn’t solve in a lifetime.  But are the AI “answers” correct?  Are AI systems just powerful

You’ve Got a Robot Lawyer in Your Pocket (Really?)

The DoNotPay AI lawyer program might be useful for fighting parking tickets but it is unsuited to serious litigation where much more complex issues are at stake
The Gutfeld! program on Fox News on January 6, 2023, recently had fun discussing robots replacing lawyers to practice law. In faux serious rhyme, Greg Gutfeld intoned: “Can a computer that’s self aware, keep you from the electric chair?” Sparking the conversation was the report that an artificial intelligence (AI) smartphone app was slated to assist a defendant fighting a parking ticket in a currently-undisclosed courtroom: Gigabytes of text could stream forth addressing the near infinite number of questions raised about robot lawyers. For now, let’s just explore the “robot lawyer” app built by DoNotPay. The company’s website declares: “The DoNotPay app is the home of the world’s first robot lawyer. Fight corporations, beat bureaucracy and sue anyone at

Defining the Role of AI in Patents

Recently, a piece of art called “Théâtre D’opéra Spatial” took home the first-place prize at the Colorado State Fair’s fine art competition in the category of digital arts/digitally manipulated photography. The art was generated using AI. Can AI hold a copyright? Can a human hold a copyright for a piece of artwork that they used AI to generate? Robert J. Marks discusses copyright, trademarks, and artificial intelligence with attorney and author, Richard W. Stevens. Additional Resources Robert J. Marks at Discovery.org Richard W. Stevens at Discovery.org The Reasonable Robot: Artificial Intelligence and the Law by Ryan Abbott  “Can a Robot Hold a Patent?” at Mind Matters News “Should AI Be Granted Patents On The Designs It

Can AI Be Issued Patents?

Should a computer program ever be listed as an inventor of a patent? Would AI have any right to sue for patent infringement? The US Patent Office has ruled that only “natural persons” can own patents, not machines, but should that change? Robert J. Marks discusses patent law and artificial intelligence with attorney and author, Richard W. Stevens. Additional Resources Robert J. Marks at Discovery.orgRichard W. Stevens at Discovery.orgThe Reasonable Robot: Artificial Intelligence and the Law by Ryan Abbott “Can a Robot Hold a Patent?” at Mind Matters News“Should AI Be Granted Patents On The Designs It Helps Develop?” at Mind Matters News“Should AI Hold Patents? The Flash-Of-Genius Answer” at Mind Matters News“Why

Patents and the Creativity Requirement

A new invention has to produce unexpected or surprising new results that were not anticipated by existing technology in order to be patented. Can computers generate something outside the explanation or expectation of the programmer? Robert J. Marks discusses patent law, creativity, and artificial intelligence with attorney and author, Richard W. Stevens. Additional Resources Robert J. Marks at Discovery.org Richard W. Stevens at Discovery.org The Reasonable Robot: Artificial Intelligence and the Law by Ryan Abbott  “Can a Robot Hold a Patent?” at Mind Matters “Should AI Be Granted Patents On The Designs It Helps Develop?” at Mind Matters “Should AI Hold Patents? The Flash-Of-Genius Answer” at Mind Matters

Law: Doe vs. GitHub Is a Non-Crisis

Despite worrisome headlines in the media, Doe v. GitHub, Inc. would protect licensed software code without blocking AI systems from using internet data for “learning”
Headline at The Verge: “The lawsuit that could rewrite the rules of AI copyright.” Wired similarly declares: “This Copyright Lawsuit Could Shape the Future of Generative AI.” The subtitle warns: “Algorithms that create art, text, and code are spreading fast — but legal challenges could throw a wrench in the works.” Indeed, two putative class action lawsuits were filed in the Northern District of California federal district court in November 2022 against GitHub, GitHub’s owner Microsoft, OpenAI and others. The lawsuits allege that two interrelated artificial intelligence (AI) software systems are continuously violating the 1998 Digital Millennium Copyright Act (DMCA) as well as breaching contracts, engaging in unlawful competition, and violating California state

Patents: A License To Sue

What is the criteria for being granted a US patent? US Citizenship? Scrutiny? How long does a patent last?

Why Don’t Robots Have Rights? A Lawyer’s Response

Robots are hardware and software packages that lack a nature or any abilities outside of whatever their designers imagine
“Free the Robots!” “Equal Rights for Robots!” Or maybe: “Set Us Robots Free!” Such future protest signs might well pop up in social media, to judge from “Why don’t robots have rights?” (Big Think, October 31, 2022) Writer Jonny Thomson worries that “ future generations will look back aghast at our behavior” when humans can “no longer exploit or mistreat advanced robots” as will presumably be the case in the 21st century. Dig into the article and get techno-whiplashed as Thomson suddenly starts talking about “the 22nd century are our friends, colleagues, and gaming partners.” Thomson’s article considers robot rights as analogous to animal rights. The summary asserts: When discussing animal rights and welfare, we often reference two ideas:

Designed to Dine, Part 2: How, Exactly, We Compute Flavor

Once a universally enjoyed but scientifically ignored phenomenon, flavor bursts out as an extraordinary event of a biological computer
Since Part 1 of this article was served up, have you experienced food and drink with greater awareness of flavor? Part 1 laid out the elements of flavor, including the smell, taste, texture, and mouth feel of foods and drinks. Smell delivers 80% of what we experience as flavor, coming from the thousands of sensory nerves in our noses detecting individual molecules. From the tongue comes taste sensations of sweet, salty, sour, bitter, and umami (savory). Flavor is like a dynamic 3-D hologram, a multi-dimensional composition of sensory inputs fluctuating in real-time, all delivered as data to the brain for final processing. Flavor makes eating fun! The Computation of Yummy The complicated and integrated systems of smell, taste, and other senses combine to deliver the experience of

Designed to Dine: Humans are Computers of Flavor

Food itself has no flavor at all. Flavor is in the sensations — really the brain — of the beholder (and taster)
Cranberries/Julie Roche, Pixabay Whether you’re a professional gourmet, a self-styled “foodie,” or an everyday North American who likes to eat, you probably look forward to celebration dinners. At any feast on Thanksgiving, Christmas, Easter, or the Passover Seder, the focal feature is the food. It doesn’t occur to us to ask: How do we sense the flavors of the food? After all, the food itself has no flavor at all. Flavor is in the mouth — and the nose, tongue, eyes, inner ears, and really the brain — of the beholder. Venture to learn how human beings enjoy food, and you’ll discover exquisite evidence of intelligent design. Like so many biological systems, detecting flavor involves specialized hardware components and the corresponding software to process information

Hawaiʻi’s Indefinite COVID Lockdown: How Would an AI Rule?

The governor of Hawaiʻi claimed that legislation supported his right to extend draconian COVID lockdown rules indefinitely. Here’s a test for an AI law program
Some people say artificial intelligence (AI) systems can become more intelligent, more intellectually capable, than humankind. After all, they say the AI “AlphaZero has taught itself chess from scratch in just a few hours and then went on to beat the world’s previous best chess-playing computer program.” AI already reads x-rays, drives cars, orders meals by phone, diagnoses skin cancer, and predicts the next movie hit. Some say AI will soon do legal analysis and make judicial decisions more accurately and fairly than humans can. Using a recent true case, let’s overview what it takes to write AI software that would analyze a statute. First, as in an appellate court brief, let’s set up the real life problem and the two contending positions. What triggered the legal

Did the Court Really Say Bees Are Fish?

And would an AI-run court — which some propose — make a different decision? Not here because California law allows the interpretation
See headlines like: “Great Day” For Bumblebees as Californian Court Rules That They Are Fish and: Bees are fish, California court rules You’d believe, on reading them, that a California court recently ruled that bees are fish. Another eyeroll-worthy court decision! Readers here might muse, “An artificial intelligence-run legal system would never make such a crazy ruling!” The Seemingly Boring Narrow Issue Let’s skip past the exciting headlines. The California Court of Appeal in Almond Alliance of California v. Fish & Game Commission faced the issue of “whether the bumble bee, a terrestrial invertebrate, falls within the definition of fish, as that term is used in the definitions of endangered species, threatened species, and candidate species” under

The Courts: May Social Media Censor Speech and Ban Users?

Two federal appeals courts came down on opposite sides. Hear the story
May Twitter, Facebook, or YouTube censor your posts and ban you from using their social media platforms? May a state government require large social media platforms to allow users and posts to present lawful information, ideas, and viewpoints with which the platforms disagree? Florida and Texas both enacted laws to restrict platforms from censoring and banning users whose content the platforms disliked. Two different federal appeals courts in 2022 ruled on whether these two states’ laws were constitutional — and came out on opposite sides. The following three scenarios frame the key issues. Scenario A: Commercial Ground Transportation A fellow boards a private company-owned, regularly scheduled commercial bus bound for Berkeley, California, wearing a T-shirt proclaiming in

Can You Trust Wikipedia to Decide Your Courtroom Fate?

Should judges and lawyers rely on Wikipedia to guide court case decisions? Researchers devised a clever test to see if they do
Wired recently ran an article entitled “Wikipedia Articles Sway Some Legal Judgments.” The subtitle declared: “An experiment shows that overworked judges turn to the crowdsourced encyclopedia for guidance when making legal decisions.” Wired’s headline oversold the story but the topic is worth a close look. Researchers at Maynooth University in Ireland, MIT, and New York’s Cornell University conducted a study to test whether Wikipedia articles about Irish court decisions affected judicial rulings in subsequent cases in Ireland. The researchers selected Irish Supreme Court decisions, analyzed them, and posted about 75 articles in Wikipedia describing those decisions. They wanted to see whether Irish courts were then using the Wikipedia articles when writing their own

US Federal court rules: Machines do not “invent” things

Evidently, Stephen Thaler’s aim was to get the patent office to recognize that an AI system can invent things all by itself
Check out this headline from lawandcrime.com: Federal Appellate Court Rules AI Systems Cannot Be “Inventors” Because They Are Not Human. Notice the angle: framing a battle between machines and homo sapiens, pitting human intelligence against artificial intelligence. The article’s first sentence spotlights the center attraction, stating: The U.S. Court of Appeals for the Federal Circuit ruled Friday that artificial intelligence or “AI” systems cannot patent their inventions because they are not “natural people.” Here the Law and Crime article subtly inserts two key beliefs: (1) that AI systems can in fact invent things all by themselves; and (2) that AI systems physically can “patent their inventions.” The sentence thus implies that the

“Do You Hear What I Hear?” Christmas Music is All in Your Mind

Is music a matter of matter and energy alone, or is there something more to the story?
What better time than the Christmas season to explore immaterial realities of the human mind? A perfect example to consider is Christmas music. It’s everywhere during the holiday season. But what exactly is music? Described in purely physical terms, music is what humans sometimes perceive from the vibrations of air. Individual pieces of music are described less in physical terms and more in subjective terms using words that reflect how humans experience music in their minds. Eight key elements of music fall mostly into the category of qualia, i.e., experiences that occur in the hearers’ minds only: Dynamics, Form, Harmony, Melody, Rhythm, Texture, Timbre and Tonality. Do you know the Christmas song “We Three Kings” when you hear

For Ants, Building a Bridge Is No “Simple” Task

There is nothing “simple” about designing neural systems and the computer systems to receive and interpret neural sensory inputs
Researching for my previous Mind Matters article about bird and bee biological software, I came across a short piece at Quanta Magazine entitled “The Simple Algorithm That Ants Use to Build Bridges.” Really, a “simple” insect algorithm? Intriguing. Eric Cassell’s book, Animal Algorithms (2021), reveals the complex and intricate hardware-software systems enabling bird and insect procedures for migration, building nests and structures, social cooperation, and navigation. Grounded in engineering training and experience, Cassell shows that animal algorithms must be designed top-down starting with a goal, fashioning the data input sensors, developing the necessary procedures, and implementing them in software to direct hardware. Yet