David Ziff
@djsziff.bsky.social
Teaching Professor, University of Washington School of Law #LegalWriting
My stuff: https://linktr.ee/djsziff
My stuff: https://linktr.ee/djsziff
Pinned
David Ziff
@djsziff.bsky.social
· Nov 11
Um. Hello new followers. I apologize in advance for my posts. I write about legal citation and #LegalWriting generally. E.g.,
A 27-page review of The Bluebook: digitalcommons.law.uw.edu/faculty-arti...
An essay on the citation of so-called "slave cases": digitalcommons.law.uw.edu/faculty-arti...
1/
A 27-page review of The Bluebook: digitalcommons.law.uw.edu/faculty-arti...
An essay on the citation of so-called "slave cases": digitalcommons.law.uw.edu/faculty-arti...
1/
Out: Using AI to write your brief.
In: Providing the judge with a link to an AI chatbot to converse about your arguments.
In: Providing the judge with a link to an AI chatbot to converse about your arguments.
New: AI chatbots can change voters' minds, according to a pair of in-depth studies published just now in Science and Nature.
How they do it is interesting — and concerning. Gift link: wapo.st/49RSstP
How they do it is interesting — and concerning. Gift link: wapo.st/49RSstP
Voters’ minds are hard to change. AI chatbots are surprisingly good at it.
New research suggests AI chatbots can shift people’s political views more effectively than campaign ads on TV.
wapo.st
December 4, 2025 at 8:32 PM
Out: Using AI to write your brief.
In: Providing the judge with a link to an AI chatbot to converse about your arguments.
In: Providing the judge with a link to an AI chatbot to converse about your arguments.
Out: Using AI to write your brief.
In: Providing the judge with a link to an AI chatbot to converse about your motion arguments.
In: Providing the judge with a link to an AI chatbot to converse about your motion arguments.
New: AI chatbots can change voters' minds, according to a pair of in-depth studies published just now in Science and Nature.
How they do it is interesting — and concerning. Gift link: wapo.st/49RSstP
How they do it is interesting — and concerning. Gift link: wapo.st/49RSstP
Voters’ minds are hard to change. AI chatbots are surprisingly good at it.
New research suggests AI chatbots can shift people’s political views more effectively than campaign ads on TV.
wapo.st
December 4, 2025 at 8:26 PM
Out: Using AI to write your brief.
In: Providing the judge with a link to an AI chatbot to converse about your motion arguments.
In: Providing the judge with a link to an AI chatbot to converse about your motion arguments.
Not the point, I know, but is this really NYT style on a number in a headline? It's "2-Decade," not "Two-Decade"?
December 4, 2025 at 6:29 PM
Not the point, I know, but is this really NYT style on a number in a headline? It's "2-Decade," not "Two-Decade"?
Reposted by David Ziff
A #legalwriting citation issue. In its certification, the fed court used Bluebook to cite an Ohio Revised Code Chapter. In its entry accepting the certification, the Ohio Supreme Court Ct switched the cite to its own style manual form w/o noting the change, but it used brackets to show a correction.
December 3, 2025 at 9:34 PM
A #legalwriting citation issue. In its certification, the fed court used Bluebook to cite an Ohio Revised Code Chapter. In its entry accepting the certification, the Ohio Supreme Court Ct switched the cite to its own style manual form w/o noting the change, but it used brackets to show a correction.
Reposted by David Ziff
"[A]n LLM does not reason in the way a litigant must. To put it in a slightly different way, LLMs do not perform the metacognitive processes that are necessary to comply with Rule 11."
This gets at a problem I have with the use of generative A.I. The process is the point!
#LegalWriting
/fin
This gets at a problem I have with the use of generative A.I. The process is the point!
#LegalWriting
/fin
December 3, 2025 at 10:20 PM
"[A]n LLM does not reason in the way a litigant must. To put it in a slightly different way, LLMs do not perform the metacognitive processes that are necessary to comply with Rule 11."
This gets at a problem I have with the use of generative A.I. The process is the point!
#LegalWriting
/fin
This gets at a problem I have with the use of generative A.I. The process is the point!
#LegalWriting
/fin
Reposted by David Ziff
This opinion from E.D. Mich. on Rule 11 and generative A.I. is an absolute *banger*. law.justia.com/cases/federa...
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
Jarrus et al v. Governor of Michigan et al, No. 4:2025cv11168 - Document 176 (E.D. Mich. 2025)
Jarrus et al v. Governor of Michigan et al, No. 4:2025cv11168 - Document 176 (E.D. Mich. 2025) case opinion from the Eastern District of Michigan U.S. Federal District Court
law.justia.com
December 3, 2025 at 10:20 PM
This opinion from E.D. Mich. on Rule 11 and generative A.I. is an absolute *banger*. law.justia.com/cases/federa...
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
Reposted by David Ziff
A lawyer's overstating of a case's holding might be permissible because the lawyer "could, perhaps, reason their way to showing how a case's stated holding might extend to novel situations." But an A.I.'s strained summary of a holding does not do that!
#LegalWriting
2/
#LegalWriting
2/
![way to showing how a case’s stated holding might extend to novel situations, an LLM does not reasonin the way a litigant must. To put it in a slightly different way, LLMs do not perform the metacognitive processes that are necessary to comply with Rule 11. LLMs are tools that “emulate the communicative function of language, not the separate and distinct cognitive process of thinking and reasoning.” Benjamin Riley, Large language mistake, The Verge https://www.theverge.com/ai-artificial-intelligence/827820/large-language-models-ai-intelligence-neuroscience-problems [https://perma.cc/7EHD-PLLZ]. When an LLM overstates a holding of a case, it is not because it made a mistake whenlogically working through how that case might represent a “nonfrivolous argument for extending, modifying, or reversing existing law or for establishing new law;” it is just piecing together a plausible-looking sentence – one whose content may or may not be true.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:o6o2u2byc2cellhzzisujort/bafkreibulxeb7xjxeu4x4vpnvjioce3v56qfxgfajy4w7qphe2incglora@jpeg)
December 3, 2025 at 10:20 PM
A lawyer's overstating of a case's holding might be permissible because the lawyer "could, perhaps, reason their way to showing how a case's stated holding might extend to novel situations." But an A.I.'s strained summary of a holding does not do that!
#LegalWriting
2/
#LegalWriting
2/
Love this opening line from a pica beats song: "Hope was not a Smith family tradition... neither was coming home sober with your paycheck still intact."
December 4, 2025 at 6:23 PM
Love this opening line from a pica beats song: "Hope was not a Smith family tradition... neither was coming home sober with your paycheck still intact."
Reposted by David Ziff
How are “metacognitive processes” different from “cognitive processes”?
"[A]n LLM does not reason in the way a litigant must. To put it in a slightly different way, LLMs do not perform the metacognitive processes that are necessary to comply with Rule 11."
This gets at a problem I have with the use of generative A.I. The process is the point!
#LegalWriting
/fin
This gets at a problem I have with the use of generative A.I. The process is the point!
#LegalWriting
/fin
December 3, 2025 at 11:36 PM
How are “metacognitive processes” different from “cognitive processes”?
Reposted by David Ziff
How many more sanctions must be ordered before it sinks in (especially to my fellow plaintiff attorneys) that any "efficiency" gained by using chatbots to write briefs is completely illusory? You are not saving time by having a machine first write a crappy brief that you MUST substantively rewrite!
This opinion from E.D. Mich. on Rule 11 and generative A.I. is an absolute *banger*. law.justia.com/cases/federa...
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
Jarrus et al v. Governor of Michigan et al, No. 4:2025cv11168 - Document 176 (E.D. Mich. 2025)
Jarrus et al v. Governor of Michigan et al, No. 4:2025cv11168 - Document 176 (E.D. Mich. 2025) case opinion from the Eastern District of Michigan U.S. Federal District Court
law.justia.com
December 3, 2025 at 10:32 PM
How many more sanctions must be ordered before it sinks in (especially to my fellow plaintiff attorneys) that any "efficiency" gained by using chatbots to write briefs is completely illusory? You are not saving time by having a machine first write a crappy brief that you MUST substantively rewrite!
This opinion from E.D. Mich. on Rule 11 and generative A.I. is an absolute *banger*. law.justia.com/cases/federa...
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
Jarrus et al v. Governor of Michigan et al, No. 4:2025cv11168 - Document 176 (E.D. Mich. 2025)
Jarrus et al v. Governor of Michigan et al, No. 4:2025cv11168 - Document 176 (E.D. Mich. 2025) case opinion from the Eastern District of Michigan U.S. Federal District Court
law.justia.com
December 3, 2025 at 10:20 PM
This opinion from E.D. Mich. on Rule 11 and generative A.I. is an absolute *banger*. law.justia.com/cases/federa...
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
The court specifically explains why the problem of A.I. is not limited to "hallucinating" case citations. Even plausible legal "analysis" *from an A.I.* is trouble!
#LegalWriting
1/
Reposted by David Ziff
fun bibliographic fact for @djsziff.bsky.social
December 2, 2025 at 8:09 PM
fun bibliographic fact for @djsziff.bsky.social
Reposted by David Ziff
And don't trust Google AI results. (The links don't actually reflect the "rule" Google provides here.)
November 30, 2025 at 6:56 PM
And don't trust Google AI results. (The links don't actually reflect the "rule" Google provides here.)
Reposted by David Ziff
My early new year's resolution* is that in 2026 everyone is going to get much better at writing, e.g., this:
What do we talk about when we talk about "the law"?
and not, e.g., this:
What do we talk about when we talk about "the law?"
*For you.
#copyediting
What do we talk about when we talk about "the law"?
and not, e.g., this:
What do we talk about when we talk about "the law?"
*For you.
#copyediting
November 30, 2025 at 6:16 PM
My early new year's resolution* is that in 2026 everyone is going to get much better at writing, e.g., this:
What do we talk about when we talk about "the law"?
and not, e.g., this:
What do we talk about when we talk about "the law?"
*For you.
#copyediting
What do we talk about when we talk about "the law"?
and not, e.g., this:
What do we talk about when we talk about "the law?"
*For you.
#copyediting
2 Honey 2 Crisp
Honeycrisp: Minnetonka Drift
Honeycrisp: Minnetonka Drift
November 29, 2025 at 11:39 PM
2 Honey 2 Crisp
Honeycrisp: Minnetonka Drift
Honeycrisp: Minnetonka Drift
Reposted by David Ziff
Update: the Routledge citation guide is making me long for the consistency and accuracy of the Bluebook
(cc: @djsziff.bsky.social)
(cc: @djsziff.bsky.social)
A dinner I am hosting got moved to tomorrow bc of snow; can somebody please tell my brain that this means i have to do citation-formatting tasks today???
November 29, 2025 at 9:42 PM
Update: the Routledge citation guide is making me long for the consistency and accuracy of the Bluebook
(cc: @djsziff.bsky.social)
(cc: @djsziff.bsky.social)
Reposted by David Ziff
What the hell is the capitalization logic at work here?
November 28, 2025 at 6:38 PM
What the hell is the capitalization logic at work here?
Reposted by David Ziff
Katie is spot on here. I recently saw a slide deck where the unrelated images took up more than half the slides and the text was illegible. Stop using garbage like Canva. Stop including all the language of the presentation on your slides. Stop reading your slides word for word. 1/2
I don’t know if anyone else notices or cares, but when I see a presentation in which the speaker uses obviously generated-AI images to illustrate their slides, it makes me immediately less confident in whatever other content they’re presenting.
November 28, 2025 at 6:38 PM
Katie is spot on here. I recently saw a slide deck where the unrelated images took up more than half the slides and the text was illegible. Stop using garbage like Canva. Stop including all the language of the presentation on your slides. Stop reading your slides word for word. 1/2
Reposted by David Ziff
I don’t know if anyone else notices or cares, but when I see a presentation in which the speaker uses obviously generated-AI images to illustrate their slides, it makes me immediately less confident in whatever other content they’re presenting.
November 28, 2025 at 3:07 PM
I don’t know if anyone else notices or cares, but when I see a presentation in which the speaker uses obviously generated-AI images to illustrate their slides, it makes me immediately less confident in whatever other content they’re presenting.
Reposted by David Ziff
ELPHABA
and
GLINDA
will return
SUMMER 2027
in SINNERS 2
and
GLINDA
will return
SUMMER 2027
in SINNERS 2
November 26, 2025 at 2:56 PM
ELPHABA
and
GLINDA
will return
SUMMER 2027
in SINNERS 2
and
GLINDA
will return
SUMMER 2027
in SINNERS 2
Reposted by David Ziff
Excellent thread on the distinction between law school memo writing and law school exam writing.
This is good 1L exam advice from Prof. Kerr. But I feel the need to provide an addendum: While this A+ answer is undoubtedly an A+ answer for a timed in-class exam in a doctrinal class, it would *not* receive a high grade in my legal writing class. Timed in-class exam answers are their own thing! 1/
Especially for the 1Ls out there, preparing for their 1st set of law school exams, I wrote this post back in 2007 about how to get a good grade on a law school issue spotter. The post, "Bad Answers, Good Answers, and Terrific Answers," is available here:
volokh.com/posts/116838...
volokh.com/posts/116838...
November 24, 2025 at 5:00 PM
Excellent thread on the distinction between law school memo writing and law school exam writing.
Reposted by David Ziff
This is good 1L exam advice from Prof. Kerr. But I feel the need to provide an addendum: While this A+ answer is undoubtedly an A+ answer for a timed in-class exam in a doctrinal class, it would *not* receive a high grade in my legal writing class. Timed in-class exam answers are their own thing! 1/
Especially for the 1Ls out there, preparing for their 1st set of law school exams, I wrote this post back in 2007 about how to get a good grade on a law school issue spotter. The post, "Bad Answers, Good Answers, and Terrific Answers," is available here:
volokh.com/posts/116838...
volokh.com/posts/116838...
November 22, 2025 at 6:36 PM
This is good 1L exam advice from Prof. Kerr. But I feel the need to provide an addendum: While this A+ answer is undoubtedly an A+ answer for a timed in-class exam in a doctrinal class, it would *not* receive a high grade in my legal writing class. Timed in-class exam answers are their own thing! 1/
Now you might think the "Downtown" on this sign was for Downtown. But no! It's for the "downtown" of a city called Federal Way, which is the opposite direction of downtown Seattle.
I do think something on this sign should tell you which direction to stand on to catch a train to downtown Seattle.
November 22, 2025 at 11:29 PM
Now you might think the "Downtown" on this sign was for Downtown. But no! It's for the "downtown" of a city called Federal Way, which is the opposite direction of downtown Seattle.
So, that said, let's look at the A+ answer, not as a professor grading an in-class timed exam, but as a legal writing assignment.
1. Ugh, that first sentence. The liability "hinges" on something? Don't tell me what liability hinges on. Just give me the answer! This isn't a mystery novel!
6/
1. Ugh, that first sentence. The liability "hinges" on something? Don't tell me what liability hinges on. Just give me the answer! This isn't a mystery novel!
6/
November 22, 2025 at 8:04 PM