I recently found myself reading Mark Lemley’s paper The Benefit of the Bargain while also helping a friend put together the Terms of Service (ToS) for their new startup. Lemley’s paper essentially argues that modern ToS - documents that are written by services to be one sided and essentially imposed on users as a take-it-or-leave-it offer - should no longer be enforced as contracts because they have lost important fairness elements of what make contracts contracts.

This argument, which I found fairly compelling, mostly focuses on the harm that the modern ToS regime does to users. ToS allow companies to impose a wide range of conditions on users that are beyond the scope of what users would ever reasonably agree to if they were offered a meaningful choice. That is in addition to the unreasonable expectation that everyday people are reading the millions of words worth of contracts they agree to in any given week.

Since I was reading this article while helping to draft ToS for a new service, I was also drawn to something the article did not mention: the ways in which these unilaterally imposed ToS hurt the other entities connected to them. Specifically, the lawyers who draft them and the companies that offer them.[1]

The Drafting Lawyers

I want to start with the most sympathetic characters in this drama: the lawyers hired to write ToS that are heavily skewed in favor of their client. Spare a thought!

It is possible to imagine such a lawyer who is pulled between two competing forces.

On one hand, they know various types of clauses in these agreements are Bad Policy, or at the least unfair to users. Such a lawyer might agree with Lemley that the world would be better without a default set of fair rules, and with a presumption that those rules could only change if the users made a meaningful choice. In their heart of hearts, they might want to draft ToS that they thought more fairly balanced the interests of the company and the company’s users.

On the other hand, that same lawyer is bound by some form of a duty to vigorously represent their client. These unbalanced terms are clearly in the company’s (at least short term) interest. Furthermore, they are essentially industry standard. As a result, this lawyer might worry that it could be a form of malpractice to fail to include the unbalanced terms in the ToS.

When faced with this tension, the lawyer might try to explain to their client that there are long term benefits to maintaining a balanced agreement with users, and therefore to leave out the most one-sided clauses. However, and there is no small irony here, it could be hard to give the client enough information so that they could meaningfully opt out of the tilted ToS arms race.

It is hard to understand the cost of giving up the short, medium, and long-term advantages provided by unbalanced ToS in service of a larger principle of social fairness. There are some startup founders who are interested in the discussion and have the bandwidth to actually process it. There are many more who will never prioritize the discussion enough to meaningfully consent to giving away the advantage.

Thus, the lawyer may face two options: a) vigorously represent their client’s interest and support the bad equilibrium by writing an industry-standard, unbalanced ToS, or 2) get out of the writing ToS business for anyone without the time and inclination to wade through the larger policy arguments.

Those feel like bad options!

The Company

This state of affairs can also harm the company, and not just in the “forcing your customers into unbalanced agreements is bad karma” kind of way.

As Lemley’s paper points out, in the offline world most businesses operate without any sort of formal written contracts at all (“you didn’t sign a contract governing the purchase of an apple from the grocery store.”). The ability to append a ToS to every digital transaction has helped to create an expectation that companies will do exactly that.

In order to do so, the companies need to take the time to write those ToS in the first place. Suddenly, instead of focusing on building and shipping their widgets, companies spend time with lawyers making sure that their ToS include all of the advantages they could possibly claim.

That’s probably a waste of time for just about everyone involved. This is made even more of a waste of time because, when faced with this new obligation, most small companies don’t hire lawyers (which would be one type of waste of resources). Instead, they tap someone without a legal background to semi-arbitrarily assemble their ToS from random corners of the internet (a slightly different type of waste of resources). And I suspect that person will increasingly outsource that task to generative AI (a third type of waste of resources).

These companies don’t really understand what unbalanced terms they are imposing on their users, what advantages they receive from them, and probably would not miss them if they were not there. They are just checking a box they don’t fully understand because it has ended up on the “things startups do” list.

I think all of these behaviors argue in favor of Lemley’s ultimate suggestion that we make a policy choice to move towards a default set of balanced rules and away from unbalanced ToS. Until then, the current system is so broken that it might make you feel bad for the lawyers and companies supposedly benefitting from it.

[1] I’ve been to enough “your paper should actually be my paper” peer reviews that I want to be clear that nothing in this post is intended to suggest that the Lemley paper is incomplete without including these points, or even that they did not occur to him. Word counts, and time, are limited in this life. Something that is interesting to me does not need to be interesting to everyone else.

hero image: a portion of Lawyers in dispute from the Met’s open access collection.

Licensing Deals Between AI Companies and Large Publishers are Probably Bad

Licensing deals between AI companies and large publishers may be bad for pretty much everyone, especially everyone who does not directly ...… Continue reading