ContractsProf Blog

What's All the Fuss About? Lawyering in the Age of AI

Adding to our extraordinarily popular "Teaching Assistants," "Reefer Brief," and "Friday Frivolity" features, I've decided to add a new "What's All the Fuss About?" series, in which we have a look at contracts scholarship that everyone is downloading.  The first in this series, avant la lettre was last year's post on Dave Hoffman and Yonathan Arbel's smash hit, Generative Interpretation.  This time, we have a look at Lawyering in the Age of AI, by Jonathan Choi (left), Amy Monahan (below, right), and Dan Schwarcz (below, left) (collectively "the Authors").

Because we are all terrified that AI will replace us, most studies of AI have tested how well various large language models (LLMs) do when asked to perform certain tasks or undertake some bit of legal analysis.  The Authors are more hopeful that AI will become a tool that humans can use to assist in legal work.  To that end, the Authors have conducted "the first randomized controlled trial of the effect of AI assistance on human legal analysis." (3) They randomly assigned sixty students a to complete various legal tasks, some with, some  without the assistance of GPT-4.  The students received a limited amount of training on using GPT-4.  The Authors then blindly graded the results. (3) The results were that the students' work improved slightly, but their efficiency improved markedly, with time savings ranging from 11.8% to 32.1%, depending on the nature of the task. (4) Based on these findings, the Authors recommend that legal educators begin thinking now about how to incorporate AI in to teaching and that, for legal practitioners, the time to embrace AI is now. (4)

The paper provides a quick overview of the interaction of technology and lawyering in the past fifty years.  Earlier innovations "improved lawyer efficiency rather than fundamentally altering the core skills needed to be an effective lawyer." (5) Early evidence suggests that LLMs might become the great equalizer.  Some studies show that they help weaker students considerably but do not really assist strong students at all. (8-9) The Authors' results provide more evidence of this pattern. (16) While the performance of AI on its own varies widely across different fields, there is already strong evidence that AI holds great promise when used to assist humans in performing intellectual work. (12) The Authors also surveyed their test subjects about their impressions of how helpful the AI was.  Their "respondents accurately perceived how useful GPT-4 was for specific tasks." (31)

Reflecting on the findings, the Authors observe that "AI tools like GPT-4 and a limited amount of training can substantially improve the efficiency with which they complete a broad array of legal tasks without adversely affecting (or even slightly improving) the quality of that work product." (31) But if anything, the Authors think their results understate the potential uses of AI for attorneys, because (1) more specialized AI tools will become available; (2) users will become more familiar with the capabilities of LLMs and use them more effectively, and (3) the rate of improvement in LLMs is so rapid. (31-34) 

In light of their robust findings, the Authors' normative conclusions include some surprises.  They start with a qualified defense of the first-year law school curriculum and suggest that use of AI should be banned or substantially limited in traditional first-year doctrinal classes so that students can develop their legal reasoning skills without being able to use technological shortcuts.  In addition, because AI enables everybody to work more efficiently but helps weaker students improve their work, while not improving the work of strong students, formative assessments should not involve use of AI.  (38) However, the Authors also advocate for the creation of courses for upper-level students that train them in the use of AI. (39)

Lawyers and law firms should likewise be exploring how best to use AI to improve efficiency and quality of work product, while also being cognizant of the dangers associated with sloppy use or misuse of AI. (39-41) Their clients should know that attorneys and law firms need to achieve some efficiency gains through the use of AI. Clients may want to bring more work in house.  However, clients also need to understand that AI will not always promote efficiencies that mean less expense.  In adversarial situations, including litigation but also contract negotiation, AI will just allow the game to be played with new layers of complexity, but the game may not end more quickly or at lesser expense. (41-42) Judges too ought to allow attorneys to use AI to assist in writing briefs in order to reduce the costs of litigation and improve access to justice. (42-43)

Breach of Contract Claims Against Elon Musk Can Proceed

Back in June, CNN's Catherine Thorbecke reported that Twitter failed to pay out bonuses to "a couple thousand" employees after being acquired by Elon Musk (represented at right).  The complaint alleged that the employees were promised bonuses at the rate of 50% of the target, and those promises were repeated after Mr. Musk's acquisition of the company.  

On Christmas day, Catherine Thorbecke posted a follow up report (I hope you gave yourself the day off for New Year's, Catherine!) noting that U.S. District Judge Vince Chhabria denied Twitter's motion to dismiss former employee Mark Schobinger's breach of contract claim.  The court "reluctantly" granted the motion to dismiss Mr. Schobinger's promissory estoppel claim, with leave to amend the complaint.  The court noted that it should be "an easy fix" and expressed some consternation at California's "seemingly pointless" pleading standards.

The breach of contract claim is pretty straightforward.  Twitter promised a bonus.  Mr. Schobinger alleged that he did everything he needed to do to qualify for the promised bonus.  Twitter didn't pay.  Perhaps as the litigation will proceed, X will eventually filed a brief saying "Twitter? I don't know her."

Friday Frivolity: Tricking AI Into Selling You a Car for $1

Chris Bakke posted the following on Twitter, which he calls X

Contracts hypo: did Chris Bakke buy a Chevy Tahoe for $1?
Real life question: if you could buy any car for $1, would it be a Chevy Tahoe?

Enrique Dans reports on Medium that Mr. Bakke achieved this result by feeding the Chevy dealer's rather primitive AI what tech people call "prompt injections."  As Mr. Dans explains, "Prompt injection is when an end user of an LLM application (or any generative AI application) gives it instructions to make it bypass those the developer of the application have provided."

TikTok Joins the Exodus from Mandatory Arbitration

Way back in 2020, Nancy Kim (left) alerted us to the sea-change already begun in the world of Terms of Service (ToS).  As Nancy reported, both Door Dash and Uber were facing thousands of arbitration claims.  Under their ToS, which provided for mandatory arbitration, the companies were obligated to pay $11 million and $18 million respectively.  Welcome to the world of mass arbitration!  A year later, Nancy posted about Amazon's decision to remove mandatory arbitration from its terms of service.

Now, as Sapna Maheshwari reports in The New York Times, TikTok has joined the party, in a really charmless manner.  It has replaced arbitration with the requirement that claims be filed in one of two California courts, and it has shortened the statute of limitations to one year from the alleged harm.  It is not clear whether TikTok can make its new terms stick.  One problem is that minors make up a huge proportion of TikTok's users, and it is not clear how TikTok could make its terms stick against people under the age of 18.  Friend of the blog Omri Ben-Shahar is quoted in the article expressing skepticism that courts would enforce significant changes to ToS posted in an e-mail or some other electronic communications.  Given the requirement that claims be made exclusively in California courts, I would think the unconscionability doctrine might also come into play.

For those interested in learning more about mass arbitration.  Georgetown Law's Maria Glover (right) is the expert.  You can find her big article on subject on the Stanford Law Review's website.  A follow-up article is available on the Washington University Law Review Website.

Friday Frivolity: The Case of the Cat Head Signature

OCU 1L Marcia Clemmons (right) shared with me this story combining two things that I love: contracts and cats.  

A Reddit user known as Bradimal posted the following to start a thread:

In 2015 I had to renew my license and thought it would be funny to do cat heads as my signature, so I did. For the past 3 years I just forgot about it unless I had to show my ID and they pointed it out.  . . .  This hasn't been a problem until today when I had to sign my mortgage papers. The signing agent looked at my ID and shook his head, he was not amused. . . . I had to sign 3 cat heads over 30 times today. 

He links to images of driver's license, along with mortgage papers. both featuring his cat signature. 

These are the driver's license cats

And these are the mortgage paper cats:

Unfortunately, the Internet is to be used with caution.  The story may not be true.  Frankly, I'm comparing the images, and I'm not convinced those are the same cats.  In fact, on the mortgage papers, with respect to the cat on the left, I'm seeing more butt cleavage than feline.

As Kim LaCapria reports on the website Truth or Fiction, there had been no independent verification of the Reddit thread.   The story went viral in 2019, but everybody just repeated the story as told.  On the original Reddit thread, Bradimal allegedly first posted an image of his driver's license, including all of his personal information.  A moderator advised him to edit the post, and he did so.  The original post cannot be retrieved, so it is impossible to know whether Bradimal really exists.  

Or is it?  According to Jonas Grinevičius and anonymous, reporting on Bored Panda, Bradimal seems to be a real person named Brad Johnsen.  Their evidence of Bradimal's existence?  They interviewed him, and he was able to identify the cats in question as Pepè De La Hoya, Sanchez Fluffington, and Eric.

That checks out.

Despite his unpleasant time with the mortgage papers, Bradimal seems to have no regrets.  He advises people to have fun with their signatures.  "It’s yours and it can be anything you want.”

China Southern Airlines Honors Tickets Sold for 10 Yuan ($1.37)!

Thanks to Wayne Barnes (below, demonstrating how one faces down an airline), we have another in our series of stories of people beating up on the airlines.  The others in this series include a couple that got a refund after being seated next to a gaseous, slobbering dog, a mother forced to sell her child into slavery when the airline would not allow the child to have a seat of her own (perhaps I exaggerate a bit), and a refund for a failed trip to Easter Island.

This time it was a simple website glitch.  For two hours, as reported on CNN, through a story provided by Reuters, China Southern advertised trips for 10, 20 or 30 yuan when they should have cost 400-500 yuan.  Passengers still had to pay airport fees, but the airlines otherwise swallowed the loss.  Perhaps this AI thing isn't all bad.

The story was reported by Sophie Yu and Casey Hall, with editing by Bernadette Baum.

Friday Frivolity: Thanks for all the Unilaterals, Elon Musk!

He may not have been able to make Twitter profitable, and he may not make it to Mars, but boy is he great at generating unilateral contract hypos!

In other Elon Musk news, I just rewatched The Incredibles with my students, as I do every year when I cover restitution.  This time I was struck by what an amazing job the Pixar folks did in creating the Syndrome character, because he is basically Elon Musk with red hair.  How did they know?

 

 

 

 

 

 

 

 

 

 

 

 

I Think I Agree with Jason Alexander: Phone Companies Should not Yada Yada Contractual Boilerlate

Thanks to OCU 1L Olivia Holder (right) for sharing with me news from the world of contracts in advertising!

T-Mobile has a new ad campaign for its Metro service.  The theme is Nada Yada Yada.  I'm not quite sure what it means.  They explain that T-Mobile offers "no contracts, no price hikes, no surprises." 

This offends me as a contracts professor.  Sure, contracts of adhesion can contain nasty terms, and the boilerplate component of them can seem like a lot of yada yada yada.  But contracts are risk allocation devices.  Without contracts, what you get is price hikes and surprises.  There is going to be a lot of yada yada yada in any contract of adhesion, and no doubt T-Mobile uses contracts of adhesion just its competitors.  But consumers need to be on notice of salient terms, and just saying there is nada yada yada does not provide consumers with those terms.

I am also offended as a Seinfeld fan.  You don't yada yada the surprising part.  You yada yada the boring part that nobody cares about.  In the relevant episode, Jerry and George (Jason Alexander) admire George's yada yadaing girlfriend because she is succinct (although I quibble with their pronunciation of that word).  But she takes things too far, whence the comedy.  Seinfeld uses this basic social norm of linguistic usage to comic effect when George's girlfiend yada yadas things, like sex and shoplifting, that one would not expect someone to yada yada. 

At the end of the episode, George realigns the usage to the norm.  Resigned to his girlfriend's foibles, George narrates the demise of their relationship -- "she went shopping for shoes for a wedding and yada yada yada, I'll see her in six t0 eight months."   He skips over the part that we now can fill in for ourselves.  T-Mobile yada yadas the part its customers need to know.  It's like they are making themselves into George's girlfriend.  She's not a normative character.

The solution is clear contractual terms and no provision that permits the wireless carrier to change terms on a rolling basis.  Well, take it from here, Jason

When a phone company tells you that there is nada yada yada, they are like George's norm-breaking girlfriend -- attractive, but perhaps a thief and probably cheating on you with her ex.

Sid DeLong on Ethical Altruism and Sam Bankman-Fried

Defective Altruism: Reflections on Bankman-Fried
Sidney DeLong

As the long hand of the criminal law closes inexorably around the neck of Sam Bankman-Fried the disgraced founder of the bankrupt crypto exchange FTX, the insatiable Meaning Machine has already begun to Draw Conclusions from his fate.

Herein are mine.  This is posted during his criminal trial and may be amended.

To some law professors, the most interesting thing in the case is that, although Bankman-Fried is not a lawyer but a graduate of MIT, both his parents, Joseph Bankman (left) and Barbara Fried (right), are each respected members of the legal academic elite, professors at Stanford Law School. Bankman is an expert in tax policy; Fried is an expert in contracts and legal ethics. She is also the co-founder of Mind the Gap, a political fundraising organization that supports the Democratic Party.  Both parents are proponents of the philosophical approach of philosopher Peter Singer, known as “effective altruism,” and thoroughly imbued their son with this philosophy. Effective Altruism is said to have motivated him and several of his associates as they built the financial empire that became FTX.

Effective altruism is a modest approach to doing good that would appeal to people across the pollical spectrum in a saner world of ideas. Briefly stated, it challenges each person to maximize their effectiveness in mitigating inequality by careful application of their resources, getting the most bang for your charitable buck. For Bankman-Fried and his friends, it meant making a lot of money and giving a lot of it away to worthy causes, while keeping enough to lead the good life. Effective altruism apparently is not overly concerned with the first step in the program (making a lot of money) and so it seems not to have mattered that Bankman-Fried’s wealth came, directly or indirectly, from speculating in the wildly fluctuating values of bitcoin and other crypto currencies.

Bankman-Fried also established FTX, a cryptocurrency exchange permitting customers to buy and sell bitcoin. In the unregulated world of crypto-currency transactions, FTX offered those who wished to deal in bitcoin a trustworthy platform on which to buy and sell bitcoin with certainty that the trades would be executed and that the customer’s funds and bitcoin holdings were safe. Because it acquired a reputation as being exceptionally trustworthy, bolstered by Bankman-Fried’s celebrity, FTX grew at a prodigious rate and was soon said to be worth several billion dollars.

With the help of funding from ethically minded millionaires, Bankman-Fried and his friends also created Alameda Research, a hedge fund firm that traded and speculated in bitcoin and other crypto currencies. As a market maker, Alameda provided liquidity to some of FTX’s customers. Alameda’s trading profits were intended to be devoted to altruistic causes. During their operations both FTX and Alameda donated millions of dollars to causes their founders deemed worthy, including Democratic Party political campaigns and causes.

Under Bankman-Fried’s control, Alameda and FTX operated closely together with only casual attention to the corporate formalities. The FTX bankruptcy trustee has accused Bankman-Fried of improperly transferring customer funds back and forth from FTX to Alameda, whence they evaporated in the crypto crash, leaving FTX customers with billions of dollars of losses from their accounts. He is also accused of commingling assets, failure to maintain proper records of transactions, and a host of other breaches of fiduciary duty to the customers and creditors of FTX. He is accused of improperly using the funds of FTX for personal gain. His defense to some of these claims is much more difficult than it would be otherwise because of grossly inadequate record-keeping.

But if these allegations are true, What Does It Mean? To trace all of Bankman-Fried’s misbehavior to his parents’ training on Effective Altruism would be unjustified. Effective Altruism is not a call for a crypto-Robin Hood to take bitcoin from the rich and give it to the poor after extracting a handsome handling fee in the process. But it is predictable that the Singer philosophy is likely to become a scapegoat in the political press, especially when the altruism appears directed to Democrats rather than Republicans. One need not be a cynic to predict that teaching Effective Altruism will be banned in Florida schools.

Although Effective Altruism does not require the forms of misbehavior of which Bankman-Fried and his friends stand accused, it seems that they did not think that it excluded it either. Perhaps Effective Altruism should be recast as Ethical Altruism, limiting the raising of funds for charitable redistribution to ethical means. Ethical Altruism would disavow robber baron philanthropy as a form of secular simony in which one’s sins can be pardoned with a sufficiently large charitable foundation, But this is only a suggestion from a non-philosopher.

If Bankman-Fried had been educated at Stanford Law School, where his parents taught, instead of or after his stint at MIT, he might have avoided his present legal plight. At a general level, he would have been taught the legal constraints and fiduciary obligations that arise when dealing with other people’s money, along with the procedural ways in which to comply with them. Regrettably, in their zeal to convert Sam to the ultimate goals of altruism, his parents apparently failed to teach the less dramatic lawyer-virtues that underlie rules that they taught their law students, including those concerning fiduciary duty, commingling, conflicts of interest, breach of trust, and the other constraints one must observe when dealing with Other People’s Money.

On a more mundane level, if he had been trained as a lawyer, he would also have been forced to come to acknowledge his own ignorance of the legal and ethical relationships that were created by his novel and complex financial empire. A less self-confident Bankman-Fried would have sought legal advice on such questions as the following: What is the exact legal relationship between FTX and people who trade bitcoin on its platform? What are FTX’s record keeping obligations regarding their trades? What am I legally permitted to do and what am I prohibited from doing with regard to customer funds and bitcoin? Am I a trustee and if so to what extent? What may I do and what am I prohibited from doing with regard to transactions between FTX and Alameda? What is the best legal vehicle to achieve my goals?

I confess that the urge to speculate as to why Bankman-Fried failed to get detailed, expert advice on these and similar questions is irresistible. To me, his tragic flaw was that old Greek standby hubris, in this case induced by his having become a twenty-something billionaire who was proclaimed a genius by powerful and wealthy people who praised him for thinking outside the box. For one who was a hero in a culture that prizes recklessness (“Move fast and break things”), it would have been uncharacteristic for a genius to run all his business plays by his lawyers. And anyway, he was assured by ethical experts that his soul was pure because his motives were unimpeachable.

So that is the Meaning I have drawn from this sad tale, as we await the completion of his criminal trial. It is a modest, law school level observation that even masters of the crypto universe must know about professional and fiduciary responsibility and that altruism must be tempered by such responsibility.

As a postscript, however, as the lesser-known details of the FTX bankruptcy sadly illustrate, Bankman-Fried may conclude from this experience that merely training lawyers about their professional responsibility may not be enough to prevent ethical lapses of the sort he committed. His counsel and several other people have accused Sullivan and Cromwell, one of the largest and most prestigious law firms in the country, of a massive conflict of interest arising from its representation of both FTX, as it advised him to file for bankruptcy, and the FTX bankruptcy trustee who later attacked those transactions. The bankruptcy court has rejected these challenges after Sullivan supplemented its disclosure.

Nevertheless, Bankman-Fried might well conclude that the Sullivan firm appears to have reasoned its way around apparent ethical barriers that would have stymied less prestigious, less sophisticated lawyers: It seems to be a professional flaw that an unconscionable fee doth make casuists of us all.

Woman Sentenced for Assisting in an Abortion, and Yes, There's a Contracts Angle

Mifepristone Molecule

Susan Rinkunas at Jezebel reports that Jessica Burgess accepted a plea deal, and a Nebraska court sentenced her to two years in prison for assisting her daughter in getting pills that helped the daughter end her pregnancy at 29 weeks.  The daughter was charged as an adult and was sentenced to ninety days in jail (she served 53) and two-years' probation.  Ms. Burgess pled guilty to tampering with human remains, false reporting, and abortion after 20 weeks’ gestation.  In exchange, prosecutors dropped two other charges. 

It is not clear how the authorities learned of the abortion, but one crucial bit of information was private messages between mother and daughter exchanged on Facebook.  As Jezebel reports,

Police obtained a warrant for Facebook messages between the mother and daughter and Facebook parent company Meta complied, providing the messages in which they allegedly discussed ending Celeste’s pregnancy with pills. 

Prior to the warrant served on Facebook, prosecutors also somehow obtained medical records revealing the daughter's pregnancy.  How they did so is not clear.  According to Jezebel, while Nebraska at the time prohibited medical professionals from performing abortions after twenty weeks, only three states, including mine, but not including Nebraska, prohibit self-managed abortion.  So, the prosecutor was investigating something that was not a crime, and Ms. Burgess pled guilty under a criminal statute that did not apply to her.

But what matters to us of course are Facebook's Terms of Service.  I recently reviewed Orin Kerr's work about Terms of Service and the Fourth Amendment.  My concern with Professor Kerr's work is that I think he downplays the ways in which users of social media sites diminish their privacy protections.  In that context, I find it striking that Facebook immediately complied with the warrant in the Nebraska abortion case.  By contrast, as Rachel Weiner reported in The Washington Post, Twitter paid $350,000 for dragging its feet in complying with a subpoena in connection with the January 6th case (above left) for Donald Trump's direct messages on the site.  In the end, even a former President has no way to protect his private messages once he shares them with a social media site.  It seems that Facebook complied with the Nebraska subpoena in a context where the illegality of the underlying conduct was much less clearly established, given that hundreds of people have already been convicted or pled to criminal charges in connection with the January 6th case.

Stay safe out there people.  If you are a woman, don't get pregnant.  Learn from the Spartan Women.  And if you do, know that your should not have any expectation of privacy if you communicate, even privately, using social media.  

District Court Dismisses Claims Brought Against Amazon as a Seller of Poison

This cases raises interesting questions about corporate responsibility, given last week's musings about what we owe to each other.  What duties does a seller owe to the public when a product it sells has become a popular option for teenagers contemplating suicide?

Two teenagers committed suicide by ingesting sodium nitrate that they bought from Loudwolf, Inc. (Loudwolf) as a third-party seller on Amazon.com.  Plaintiffs, the parents of the teens, alleged that Amazon had been aware that teens were using sodium nitrate to commit suicide since 2018, but it only stopped allowing the sale of the product on its website in December 2022.  Plaintiffs sued Amazon and Loudwolf, alleging causes of action sounding in negligence and strict liability, but they later agreed to dismiss Loudwolf from the case.  In McCarthy v. Amazon.com, the U.S. District Court for the Western District of Washington granted Amazon's motion to dismiss.

In order for Amazon to be strictly liable it must count as a "seller."  Amazon disputes that it is a seller, and its status is on that count is interesting and tricky, as we learned from Tanya Monestier's scholarship, but for the purposes of its motion to dismiss, Amazon concedes that plaintiffs have alleged that it is a seller.  Under the relevant Washington Product Liability Act (WPLA), a seller can only be liable if it (1) was negligent, (2) breaches an express warranty, or (3) intentionally misrepresents facts or intentionally conceals information.

Given sodium nitrate's obvious dangerousness and its adequate warning labels, the court rejected plaintiffs' arguments that Amazon was negligent.  Moreover, Amazon's alleged failure to provide better notice of the dangers of sodium nitrate were not the proximate cause of the two teens' suicides.  They used the product precisely because they knew the risks associated with it.

Plaintiffs' argument that Amazon engaged in fraudulent concealment sends the court in a highly interesting direction.  Their claims is that Amazon is liable because it concealed or deleted negative reviews that warned of the products use for suicide.  Amazon did so because it has a general policy of removing reviews that mention suicide.  As the court explains, plaintiffs' intentional concealment claim is barred by the ever-unpopular-and-yet still-standing § 230(c)(1) of the Communications Decency Act.

The court found that both plaintiffs' common-law negligence and their negligent infliction of emotional distress claims were precluded by the WPLA.  Moreover, as the goods were not defective, plaintiffs cannot make out a negligence claim.  And given futility, the court denied leave to amend (which plaintiffs did not even seek).  

I Just Discovered the Common Law Podcast! Have You?

Somehow I had not noticed the UVA Law has a podcast called Common Law.  As near as I can tell, there is some variation but also some continuity in the hosts, and the subject matter is not always the common law.  Rather, it is a law talk blog in which the UVA hosts interview UVA and non-UVA faculty about recent legal issues and recent legal scholarship.

A recent episode, involving friend of the Blog Kim Krawiec on her favorite topic, Taboo Trades.  Inattentive as I am, I learned from the podcast for the first time the intriguing origins of Kim's interest in taboo trades.  She was working at a BigLaw firm on derivatives transactions and learned that such transactions were once regarded as a form of gambling and were illegal, as life insurance contracts once were.  I know.  Amazing.  Gambling was once illegal!

This episode is hosted by UVA's Dean Risa Goluboff and Professor Cathy Hwang.

Listen!  And lucky you!  There are five seasons that you have not yet listened to!

Bitcoin Mining Company Bound By Arbitration Agreement It Never Signed

Faes & Co. (Faes), a UK company that engages in Bitcoin mining, contracted to purchase fifty Bitcoin mining machines from a U.S. company, Blockware Solutions, LLC (Blockware) for $525,000.  The parties' relationship was governed by two agreements: a “Services Agreement” and “Co-Location Agreement.”   The Services Agreement basically just provided for Blockware's obligation to procure mining systems for Faes once Faes had purchased them.  The Co-Location Agreement covered Blockware's obligation to provide for the installation, maintenance, and operation of mining equipment, as well as electrical power and Internet.  The Co-Location Agreement included an arbitration clause.

The machines didn't work right.  I won't bore you with the technical details, and because of my limited understanding of these things, I couldn't if I tried.  It seems like the machines did not operate at full capacity.  There was a lot of down time, which cost Faes $5000/month.  So Faes sued, and Blockware dutifully moved to dismiss for improper venue, which the court generously construed as a motion to compel arbitration.  

Bitcoin Mining, image by DALL-E

In Faes & Co., Ltd. v. Blockware Solutions, LLC, the District Court for the Northern District of Illinois granted the motion to compel arbitration.  Faes argued that it was not bound by the Co-Location Agreement's arbitration clause because it never signed it.  That's just silly.  It accepted the machines and used them.  Their conduct rendered their signature unnecessary.  If Faes did not think it was contractually bound, why did it pay in exact accordance with the terms of the agreement?  Faes next argued that it was not bound by the Co-Location Agreement because it expressly rejected it.  In fact, Faes had quibbled with one clause of the agreement, but upon receiving assurances from Blackware that the clause would not operate as Faes feared, it proceeded with the transaction.  It never objected to arbitration.

The case is stayed pending arbitration.

Orin Kerr on Terms of Service and Fourth Amendment Rights (Once More, with Feeling)

I flagged Orin Kerr's Terms of Service and Forth Amendment Rights as something I intended to read and write about back in February.  I wasn't sure when I was going to get back to it, and today is the day.  The article is now forthcoming in the University of Pennsylvania Law Review (congratulations all around!), but if you can't wait until then, you can find the draft on SSRN.

The article explores the possibility that Terms of Service (ToS) might allow large corporations to turn your personal data over to the government, thus providing the government with a ready alternative to the Fourth Amendment's warrant requirement.  As Professor Kerr puts it,  "What the Supreme Court has given, Terms of Service might take away."  

Courts are divided.  Some will not allow people to contract away their rights by agreeing to ToS; most treat use of the service as a waiver of Fourth Amendment rights.  Professor Kerr argues that the courts that treat ToS as a waiver get things wrong.  He looks at four doctrines where one might think that ToS might lead to lost privacy and concludes that none of them should affect Fourth Amendment rights.   An essential step towards this conclusion is Professor Kerr's insistence that ToS clarify relationships relevant to Fourth Amendment doctrines, but those relationships, not particular terms, govern Fourth Amendment rights. 

Terms of Service, Image by DALL-E

Professor Kerr analogizes breach of ToS to breach of a car- or apartment- rental agreement.  A user's breach of those agreements has no impact on Fourth-Amendment rights.  However, he acknowledges, ToS invite a difference line of reasoning: Fourth Amendment protections require rights; ToS define rights in online accounts; therefore, ToS define Fourth Amendment rights in online accounts.  Actually there are two possible routes to the same conclusion, Professor Kerr notes.  In addition to treating contracts rights as a waiver, in cases in which a private company voluntarily turns over incriminating evidence to the governments, courts sometimes avail themselves of the third-party doctrine.  The search was not conducted by the government, and the Fourth Amendment does not protect against private searches.

Professor Kerr thinks that courts that think ToS largely irrelevant to the Fourth Amendment analysis get things right.  The law has never recognized a diminishment of the reasonable expectation of privacy just because phone companies have access to user data, or landlords or cleaners have access to an apartment.  The tough cases arise where the ToS have language clearly indicating that the provider will share incriminating information with the government.  To some courts, such language renders the ToS distinguishable from a typical lease of a car, a hotel room, or an apartment.  

Professor Kerr then makes four compelling arguments for why existing caselaw forecloses arguments that would permit the government to use ToS to get around the warrant requirement.  First, the caselaw on shared spaces provides that agreeing to share an office or an apartment with others does not deprive one of a reasonable expectation of privacy vis a vis the government.  Second, formalizing rules of access to private spaces, as one might do in a hotel, business, or apartment rental agreement does nothing to change one's expectation of privacy against the government.  Third, this is so, even if the user violates a rental agreement.  So, for example, police cannot carry out an illegal search of rental car driven by an unauthorized driver even if that amounts to a breach of the car rental agreement.  The same is true in cases involving breaches of apartment leases and use of hotel rooms.  Finally, the rule for shared space operates differently in government spaces, but Professor Kerr notes, that exception to the expectation of privacy has always been limited to government offices.

Professor Kerr then makes a remarkable discovery: the cases that treat agreement to ToS as carving out an exception to traditional Fourth Amendment doctrine rely on the government spaces cases!  They mistake the exception for the general rule.  

Finally, Professor Kerr discusses four doctrines where one might think that ToS could make a difference.  They are the private search doctrine, third-party consent, direct consent, and abandonment.  Professor Kerr thinks courts should treat ToS as irrelevant to all four doctrines.  

Here, I must admit, the argument loses me.  Either I don't understand it, or it gives me little comfort.  With respect to private search doctrine, the case on which Professor Kerr relies, United States v. Rozenow, seems to allow Internet Service Providers (ISPs) to share information with the government when their motivation for investigating an account is violation of their ToS.  Rozenow was treated as having no expectation of privacy in the material on his account because it was discovered through a private search motivated by his ISP's independent interest.  It sure seems like the ToS very much matter in determining what actions an ISP might take, beyond suspending an account, including storing incriminating material or sharing it with law enforcement.

In discussing third-party consent, Professor Kerr acknowledges that ToS might be written so as to track the doctrine and allow private companies to share incriminating data or information with law enforcement.  However, he argues that the inquiry hinges not on theToS but on the reality of shared access to a user's data.  If the ISP doesn't actually access the information, the expectation of privacy survives.  I would argue the opposite.  Companies put language in their ToS about shared access to data because they use our data.  That's whey they don't charge us; we pay by sharing our data with ISPs.  Our data could not be shared with law enforcement if there were no ToS evidencing the users' consent for ISP access of and use of the data.  It seems to follow that in many contexts, the third-party consent doctrine might provide a way around the Fourth Amendment's warrant requirement.

In discussing direct consent, Professor Kerr argues first that consent to ToS cannot be construed as consent to a government search.  The connection is too attenuated.  I am happy to defer to Professor Kerr's vastly superior knowledge of Fourth Amendment law here, and I find it entirely believable that a more stringent definition of consent should apply in the Fourth Amendment context than in the contractual context.  His second argument is less convincing: ToS cannot bind users because nobody reads them.  Isn't it pretty to think so?

Finally, his argument on abandonment references his earlier arguments.  He makes the strong point that it is hard to construe registration of an account as an abandonment of the expectation of privacy in that account.  That might be convincing but for the ToS that might provide that by using an account you are agreeing to share your data with the ISP and abandoning any expectation of privacy in that data.

Despite these critical comments, I hope that Professor Kerr's way of thinking about these things succeeds.  Nancy Kim and I expressed our concerns about Internet Giants as quasi-governmental actors nearly a decade ago.  We raised some of the same concerns about consent that Professor Kerr does.  But we adopted a pessimistic tone.  Professor Kerr's work could provide more grounds for optimism.

More generally, I have been posting in this space periodically about interactions between contracts and the First Amendment.  My view of the caselaw is that courts in certain contexts have allowed people to contract away Fourth Amendment protections, but they sometimes treat First Amendment protections as ironclad, even when the expression at issue is not core political speech, notwithstanding contractual interests.  Professor Kerr's approach would bring the jurisprudence on the relationship of contracts and the Fourth Amendment more in line with that on the relationship of contracts and the First Amendment.   

Advice to 1Ls: Course Correction

I was listening to one of my favorite podcasts this morning, the National Security Law podcast, Episode 241, "Let's Barbie."   The hosts,  Bobby Chesney (below left) and Steve Vladeck (below right), started with some words of advice for new law students.  I have my own version of this that I deliver to our 1Ls at OCU Law, and I was struck by the near-complete lack of overlap between when they said and what I say.  These two are more than merely two professors with a podcast (and a Twitter account).  They are serious scholars (see, and read! e.g., Steve's new book on the Supreme Court's Shadow Docket) who are committed to their institutions (Bobby is now Dean Bobby, but both evidence their commitment to teaching in every podcast).  

So, they started with wise words about self-care: Check in weekly with family and close friends, they advise.  Don't disappear into the four-month black hole of the first semester of law school.  And then, they advised students not to panic if it seems like other students seem to be grasping the material more quickly.  They are all there for a reason.  Things tend to even out over the course of the first semester.  This is great advice, and I said very little of it.  That was an error, but I come by it honestly.  We teach at different institutions, and I have different concerns about the dangers my students face.

I find that the students who get in the biggest trouble are students who do not realize, or whose lives do not allow them to address, how much more demanding law school is than college.  I worry, because I have experienced this, about students whose parents think they can rely on them as back-up childcare providers or miss class to go for airport runs to pick up out-of-town relatives who are coming in for an anniversary celebration.  I worry about students who live at home or are married with young children and cannot carve out daily a quiet space for extended periods of intense study.   So my first bit of advice to my students is that they need to have frank conversations with family and close friends.  They need to explain that they are not "just a student."  They have a full time job and then some.  Succeeding in law school will take their full time and attention, and family and friends need to accommodate that.  My students who do poorly in law school do not lack intelligence or drive.  They lack the time and freedom to give themselves the opportunity to succeed.

But listening to Steve and Bobby, I realize I need to temper my message.  Law students need not be ascetics.  Self-care is also important to student success.  Fortunately, I have great colleagues, including staff, whom I trust communicated that message to our entering class.   Nonetheless, I am sometimes told that my students for some reason take me very seriously.  I need to let my students know that they can allow themselves some down time and enjoy the benefits of friends and family for support and companionship.  They need not feel guilty about that or think that they would thereby be disappointing my expectations.

Weekend Frivolity: A Contracts Hypo from the X-Man (Formerly the Chief Twit)

This year, I think I may spin out some unilateral contract hypos with the help of this tweet

We can start by asking whether or not it is a real offer.  Is it sufficiently clear and definite?  I suppose we might wonder what it means to be "unfairly treated" by one's employer and who gets to determine what treatment counts as "unfair."  To whom is the tweet addressed?  Who is "we"?  Is it Mr. Musk, using the royal "we"?  Is it Twitter (no, I'm not going to call it "X")?  The use of "we" makes it seem like it is the corporation that is making the offer, but doesn't the corporation have its own Twitter account?  No limit?  Really?  At what point and through what means could it be effectively withdrawn?

Then we can imagine some scenarios . . . .

• Can a party accept this offer if they worked for a private business and were disciplined when their employers discovered that they were behind a notorious website that made fun of the business and cost it millions in good will losses?  What if they weren't formally disciplined but everybody was cool and distant?
• Can a party accept this offer if they had a government job and were disciplined when it was discovered that they were aggressively posting criticisms of the government under a pseudonym?  
• Can a party accept this offer if they work for Tesla, SpaceX, or post-Musk Twitter?
• Can a party accept this offer if they had signed a non-disparagement clause that specifically prohibited posting anything on social media critical of the employer?
• What if your employment is at will?  What if the "unfair treatment" is that your contract was not renewed?

Evidentiary issues are likely to arise.  Would an employer be so unwary of potential First Amendment issues that it would say something along the lines of "we are disciplining you because of what we found on your social media accounts?"  Even if there were some discussion of social media accounts, couldn't an employer always claim that there were other grounds for the discipline?  So who has the burden of establishing, for the purposes of the offer that the "unfair treatment" was "due to"?

Finally, or perhaps this comes first, there is no cause of action for being treated unfairly by one's employer.  I have been an employee in many settings.  From time to time, in every setting, I have felt that I was treated unfairly by my employer, as did many of my peers.  Perhaps I have a very well-developed sensitivity to unfair treatment.  Nevertheless, in none of those situations did I choose to bring legal action against my employer.  Mr. Musk's tweet can be a game-changer for people like me.

So what if I contacted Mr. Musk and accepted his offer.  No, I was not planning on initiating a legal action against my employer.  I'm not sure what my cause of action would be, but I do post on social media in ways that my employer doesn't like, and as a result, I think people shun me, and the bosses pass me over for promotions.   Is Mr. Musk obligated to assemble a legal team and fund my legal bill without limit?

Sanctions and Warnings for Attorneys' Use of ChatGPT

Emojis are pretty old technology, and yet, as we discussed last week, the law is just wrestling with whether assent and signature by emoji is effective, and so far, only Saskatchewan has weighed in on the issue.  The rest of the common-law world is sprinting to catch up.  

And so it is with the new AI chatbot technology.  Back in June, as Sara Merken reported in Reuters, two New York attorneys were disciplined for relying on ChatGPT to write a brief and not noticing that the bot had hallucinated cases.  Apparently, the lawyers failed to acknowledge the error in a timely way when opposing counsel and the court could not find the cases.   Although the court found that the attorneys had acted in bad faith and had made false and misleading statements, the court also expressed understanding that attorneys might make legitimate use of chatbots, subject to attorney gatekeeping "to ensure the accuracy of their filings." 

Benjamin Weiser, writing in The New York Times, digs deeper, covering the sanctions hearing in detail.  The attorney who was led astray heard about ChatGPT from his college-aged children but had never worked with it before.  "I used this to write all my papers this semester, Dad, and look: 2.73 GPA!"  He never imagined that it would make up cases, and why would he?  Never trust technology recommended by anyone under thirty.  

Meanwhile, other courts are being proactive, as Jacqueline Thomsen reports for Reuters, U.S. District Judge Brantley Starr of the Northern District of Texas issued at order putting attorneys on notice "that they can't just trust those databases. They've got to actually verify it themselves through a traditional database."  Other courts require attorneys to disclose when they use AI in their research.  Eugene  Volokh reports on his beloved eponymous conspiracy blog that at least one judge has prohibited attorneys admitted pro hac vice from using AI.   

That seems a bit harsh.  ChatGPT is a tool like any other research tool.  Prohibiting use of ChatGPT makes as much as sense as prohibition of use of Google or Wikipedia.  Except that those two tools have been around for a while,  But then again, I am not worried that either of them will become sentient and destroy humanity.

Human Beings Are Still Better than Bots

Two weeks ago, I reported on a decision by the National Eating Disorders Association (NEDA) to replace its Helpline, staffed by actual human beings, including six recently-unionized workers, with an AI bot named Tessa. 

Tessa has been unplugged.

As Chris Morris reports in Fortune, NEDA took down the bot after receiving complaints that it was responding in ways that promote eating disorders.  NEDA, which really needs to do some soul-searching, first responded by claiming on social media that the reports were a lie.  However, once it saw evidence of the "lie" in the form of screenshots, it shut down the bot and launched an investigation -- or so it says.  I'll have to see screenshots to believe it.  

In related news, today I beat the Wordle Bot.  It's a sore loser and protests that my third guess was lucky.  Hey, Wordle Bot, you have ONE JOB!

Guest Post from Guy Rub on Federal Law and Browsewrap Agreements

Today, we welcome guest-blogger Guy Rub (right) to the blog.  Professor Guy A. Rub is the Joanne Wharton Murphy/Class of 1965 and 1973 Professorship in Law at The Ohio State University Moritz College of Law. He is an expert in the intersection between intellectual property law, contract law, and economic theory. His work explores how markets shape and are being shaped by intellectual property law. His publications have appeared in the Chicago Law Review, UCLA Law Review, Yale Law Journal Forum, and Virginia Law Review, among others.

Professor Rub has studied law on three continents. He holds an SJD degree and an LL.M. degree from the University of Michigan Law School; a master’s degree in Law & Economics from the University of Madrid; a European Master in Law and Economics from the Erasmus University in Rotterdam, Netherlands; and an LL.B. degree from Tel-Aviv University.

Professor Rub also holds a bachelor’s degree in computer science from Tel-Aviv University and worked as a software programmer and engineer prior to pursuing a career in law.

Browsewrap Agreements and Federal Law

Standard form agreements have been a topic of high interest and controversy for decades. Still, in the abstract, most would agree that while important differences exist between negotiated contracts and form agreements, most courts most of the time treat them quite similarly. A few doctrines, like unconscionability, operate differently when form agreements are concerned, but, for most courts, once Alice signs the dotted line, a bilateral contract is formed, whether she spends ten months negotiating the terms or ten seconds skipping them. And, as we know, signing is overrated too. If Alice can accept a contract by signing a form she didn’t read, she can also do it by tearing a wrap or clicking on “I accept.” 

But browsewrap agreements, contracts that are allegedly accepted by using a website, are trickier. To the degree that the offer and acceptance ceremony still matters (and it should, right?), most people probably perceive the act of signing a dotted line, clicking “I agree,” and browsing quite differently (although, for what it’s worth, over the last decade, more and more students are telling me that “everybody knows that by using a website you accept its terms.”). So are browsewrap forms binding? It depends a little on the jurisdiction and a lot on the website. Courts have developed tests that focus on what users actually knew and on how “conspicuous” the terms were. They explore the color of the link, where it was placed, how big the font was, and so on. Browsewrap agreements let judges play website designers.

While state law still struggles with browsewrap agreements, a new brief from the United States Solicitor General (left) suggests that federal law might have something to say about the issue. Genius.com is a popular website for music lyrics. Its browsewrap prohibits the copying and public display of those lyrics for commercial use. Google, however, allegedly did just that. If you google “the lyrics of little lies” (useful when teaching misrepresentation), you will get the results on Google itself, in what Google calls “an information box.” No need to spend time visiting Genius’s website for lyrics anymore. This can make the users happy and Google probably happier, but it is a disaster for Genius’s ad-centric business model. A breach of contract lawsuit followed.  

Contract law might give Google quite a few tools to fight this lawsuit, but it chose federal preemption as its first line of defense. Google argued that Genius’s claim is expressly preempted by copyright law. Section 301(a) of the federal Copyright Act preempted rights under state law that are “equivalent” to rights under copyright. Courts disagree on whether contracts are ever equivalent to copyright. A minority of courts hold that contracts that regulate what copyright regulates, mostly the reproduction and distribution of information goods, are preempted. The majority of courts, however, hold that a contract, as a bilateral agreement, cannot be equivalent to a property right like copyright. The most famous articulation of that majority approach came in an opinion we know for other reasons—ProCD v. Zeidenberg. In Genius v. Google, however, the Second Circuit adopted the minority approach and held Genius’s browsewrap agreement preempted as a contract that regulates copying. Genius’s cert petition to the Supreme Court followed, and the Court invited the Solicitor General to file a brief on this matter. 

In its brief, the Solicitor General acknowledges the circuit split and suggests that the Second Circuit approach is quite problematic (a view I share), but it nevertheless recommends denying the cert petition. At the heart of its brief, the Solicitor General argues that this case is not a good vehicle to address the copyright-contracts tension because browsewrap agreements are just different. 

The Solicitor General claims that it is not clear that this case would have come out differently in any other circuit. It, quite creatively, partly relies on a much-quoted (and often misunderstood) sentence from ProCD, where Judge Easterbrook (right) explains that not everything “with the label ‘contract’ is necessarily outside the preemption clause.” I always assumed that if the Seventh Circuit had any specific example in mind, it was probably implied-in-law contracts, which, we know, are not contracts at all. But can’t we apply the same approach to browsewrap agreements? After all, if the heart of the Seventh Circuit argument is that the act of acceptance categorically distinguishes contracts from copyright, then maybe it is less than obvious that browsewrap agreements should be treated similarly.

To be clear, in this case, courts addressed the federal preemption question first and never decided whether a contract was even formed under New York law. But the Solicitor General suggests that even if a contract was formed, for preemption purposes, that contract “is quite different from the paradigmatic bargained-for exchange.” 

Most of the implications of this brief and this case are beyond the scope of contract law, but contract professors will still find it interesting. Consider, for example, the following nuggets: First, while its argument is interesting, the Solicitor General’s brief completely ignores the fact that Genius repeatedly put Google on actual notice as to its terms and conditions, which is likely quite meaningful as a matter of state contract law. Shouldn’t it matter for federal law too? Second, contract professors might find it noteworthy that the Solicitor General distinguishes ProCD because there “the existence of a contract was apparent.” Third, the brief not only argues that browsewrap agreements are different, which they might be, but it implies that they are maybe less important, referring to them repeatedly as “atypical.” I’m not sure that claim is self-evident. 

In any event, are we going to soon have the Supreme Court weighing in and telling us whether browsewrap terms and conditions are contracts, at least from a federal preemption perspective? I know better than to bet on anything that the Supreme Court does, especially on cert petitions. However, it is probably more likely than not that the Court will deny the petition. But if it surprises us, maybe Justice Sotomayor will get a chance to revisit the topic 20-something years after Specht v. Netscape.

AI Assists in Union Busting

As we noted last August, 2022 saw a dramatic uptick in unionization in the United States.  Not surprisingly, for every action, there is a dastardly corporate reaction.  But this is pretty extraordinary.  

According to Britney Nguyen, writing in The Business Insider, the National Eating Disorder Association (NEDA) is shutting down its human-run helpline and replacing it with an artificial intelligence program called Tessa.  Staff members who worked on the hotline were recognized as part of a labor union on March 17th.  Two weeks later, NEDA announced that it was firing its employees and replacing them with a chatbot, effective June 1st.  The human hotline was staffed by six paid workers and 200 volunteers.  If you call it now, you receive an automated message that advises people in need of human contact to use the Crisis Text Line. 

Kate Wells of National Public Radio reported that nearly 70,000 people called the Helpline in 2022, double the number who called before the COVID pandemic.  NEDA notes that the Helpline is overburdened, and people often have to wait before they can talk to someone.  A NEDA spokesperson noted that the Helpline was staffed with volunteers who were not trained to deal with crisis situations. 

Guess what the unionizing workers wanted.  More money?  No.  More training and more staff!  

The creator of Tessa, Ellen Fitzsimmons-Craft, speaking to NPR, acknowledged that the program is not capable of the kind of empathy that comes from an actual human being, but it beats just being put on a waitlist.  Okay.  That sounds like there could be a role for Tessa.  Still, Tessa is no panacea.  According to Dr. Fitzsimmons-Craft, Tessa is "really a tool in its current form that's going to help you learn and use some strategies to address your disordered eating and your body image."  But she does not represent Tessa as a substitute for the helpline.  "It's not an open-ended tool for you to talk to and feel like you're just going to have access to kind of a listening ear, maybe like the helpline was."

But I'm no expert on the AI potential for helping people with eating disorders, so I asked my good friend ChatGPT whether is it a good idea to replace an eating disorder helpline with AI.

Here is ChatGPT's response

Replacing an eating disorder helpline with AI alone may not be a comprehensive or effective solution. While AI can provide certain benefits in the field of mental health, it is important to consider the limitations and potential risks involved.

Here are a few points to consider:

1. Emotional support: Eating disorders are complex mental health issues that often require empathy, understanding, and emotional support. Human helpline operators, who are trained professionals, can provide a compassionate and empathetic response, which is crucial for individuals struggling with eating disorders. AI, on the other hand, may lack the emotional intelligence and sensitivity needed to address the nuanced needs of those seeking help.

2. Personalization: Each individual's experience with an eating disorder is unique. Human helpline operators can tailor their responses to the specific needs of the caller, providing personalized advice and guidance. AI, while capable of processing large amounts of data, may struggle to offer the same level of personalization and individualized care.

3. Ethical considerations: Privacy and confidentiality are of utmost importance when dealing with mental health issues. AI systems need to ensure the protection of personal information and adhere to ethical guidelines. However, concerns about data security and privacy breaches remain, potentially leading to trust issues and reluctance among individuals to seek help.

4. Diagnostic accuracy: Eating disorders require careful assessment and diagnosis. Qualified professionals are trained to recognize the signs and symptoms, conduct assessments, and make appropriate referrals. While AI can assist in providing information and resources, it may not possess the same level of accuracy and clinical judgment as human experts.

However, it's worth noting that incorporating AI into existing helpline services can be beneficial. AI can augment human operators by providing additional resources, information, and tools that can enhance the quality and efficiency of the service. For example, AI can help automate administrative tasks, offer educational materials, or provide self-help resources.

In conclusion, while AI can play a supportive role in eating disorder helplines, completely replacing human operators with AI may not be ideal. A combination of human expertise and AI technology can provide a more comprehensive and effective solution, ensuring that individuals receive the necessary emotional support and personalized care.