prior probability

Or, if you prefer, here is a more comprehensive explanation of large language models.

That is, will OpenAI and its ilk get sued out of existence? I certainly hope so! For your edification and possible amusement, below are some “A.I. legal risk/legal liability links”:

1. Potential legal liability for copyright infringement (via the Wall Street Journal).

2. Potential legal liability for defamation (via The Verge).

3. Potential legal liability for products liability (via SSRN).

Bonus link: Database of all lawsuits involving A.I., via George Washington University.

I could not go all the way to San Diego without visiting my favorite place in the world: my beloved Mexico! As an added bonus, my son Kleber is in town this weekend, so we crossed the border together. Memo to the U.S. Customs and Border Patrol (CBP): you suck. The two-hour line to cross back into “the land of the free” was of Soviet Russia proportions!

I have been in “America’s finest city” this week attending the annual meeting of the Academy of Legal Studies in Business (my penultimate trip this summer). Below is a small sample of some of my San Diego snapshots:

Are Hollywood and Broadway the last bastions of sexism? The Bechdel test asks whether a play, motion picture, or TV show features at least two women who talk to each other about something other than a man. Although this now-iconic litmus test began as a joke (see screenshots below), it is still rather remarkable how many–or should I say how few?–theatrical and cinematic productions are able to pass it.

To commemorate the tenth anniversary of the publication of my 2013 law review article “Gödel’s Loophole” (my most downloaded paper by far), I just posted to SSRN a new work-in-progress titled “Gödel’s Loophole 2.0“, which is now available here. (I will be presenting this paper at the annual meeting of the Academy of Legal Studies in Business this week.) In summary, my previous work from ten years ago conjectured that Gödel’s loophole in the domain of constitutional law refers to constitutional self-reference, i.e. an amendment to a constitution’s amendment clause.

My new article extends the general logic of Gödel’s loophole to two new domains. One is the field of artificial intelligence and a new proposal known as “constitutional AI”, a self-regulating system for minimizing the harmful effects of large language models like ChatGPT, Claude, and DeepMind. (See here, for example.) The other domain is the planet Mars. Although as of this writing the possibility of permanent human settlements on Mars is still far off, several draft constitutions of various lengths and levels of complexity have already been proposed for Mars! (See here, here, and here.) Alas, neither “constitutional AI” nor the sundry Martian constitutions are immune from the inexorable logic of Gödel’s loophole.