IN THE SUPERIOR COURT OF THE STATE OF DELAWARE
ROBERT REED, an individual, )
)
Plaintiff, )
)
v. ) C.A. No. N25C-10-211 MAA
)
OPENAI LLC, a limited liability )
company, )
)
Defendant. )
)
Submitted: August 2, 2026
Decided: August 24, 2026
Reed’s Motion to Dismiss:
DENIED.
OPINION
Andrew E. Russell, Esquire and Virginia Lynch, Esquire, of SHAW KELLER LLP,
Wilmington, DE, Krista L. Baughman, Esquire (Argued), of BAUGHMAN LAW,
Walnut Creek, CA, Jesse Franklin-Murdock, Esquire, of DHILLON LAW GROUP
Inc., San Francisco, CA, Attorneys for Plaintiff.
Andrew D. Cordo, Esquire, Kaitlin E. Maloney, Esquire, and Brooke D. Pridham,
Esquire, of WILSON SONSINI GOODRICH & ROSATI, P.C., Wilmington, DE,
Matthew A. Macdonald, Esquire (Argued), of WILSON SONSINI GOODRICH &
ROSATI, P.C., Los Angeles, CA, Attorneys for Defendant.
Adams, J.
INTRODUCTION
In this action, Plaintiff Robert Reed (“Reed”) sued Defendant OpenAI,
LLC (“OpenAI”) for defamation, alleging OpenAI’s artificial intelligence tools
produced defamatory statements about Reed. OpenAI moves to dismiss
Reed’s Complaint, contending Reed fails to satisfy certain necessary
elements of a defamation cause of action.
The issues presented in this case reveal a new frontier for defamation law, in
which artificial intelligence tools are allegedly employed to effectuate the
defamatory ends of their makers. This motion does not force the Court to take a
deep dive into those novel issues. Instead, the Court resolves OpenAI’s motion to
dismiss based on established defamation caselaw and the plaintiff-friendly standard
at the pleading stage. This Opinion DENIES OpenAI’s motion in its entirety.
FACTS AND PROCEDURAL HISTORY
The factual background outlined herein is drawn from the Complaint,1
accepting all well-pled allegations as true only for purposes of this Motion.2 The
Court will not necessarily use terms like “alleged” throughout. The Court intends to
convey no agreement with the truth of the matters asserted in the Complaint. The
veracity of the Complaint’s allegations can be resolved after discovery.
1
D.I. 1.
2
Cent. Mortg. Co. v. Morgan Stanley Mortg. Cap. Hldgs. LLC, 27 A.3d 531, 536 (Del. 2011)
(citation omitted).
1
A. The Parties.
Reed is a resident of Tennessee3 and is a filmmaker, journalist, and
activist.4
OpenAI is a Delaware LLC with a principal place of business in California.5
In 2017, OpenAI created the OpenAI AI division, which is focused on developing
artificial intelligence (“AI”) products.6
B. Reed identifies false statements by Bard in 2023.
In March 2023, OpenAI released Bard, its first AI product powered by large
language models (“LLM”), for early access.7 LLM-based AI products “process
natural language to provide intelligent responses in a chat, including follow-up
responses that mimic human conversation in a sophisticated fashion.”8
In December of 2023, Reed learned Bard was providing false information
regarding Reed, incorrectly tying Reed to a “notorious white supremacist.”9
Reed further learned Bard would, when asked, provide argument in favor of
3
Compl. ¶ 14.
4
Id. ¶ 2.
5
Id. ¶ 15.
6
Id. ¶ 18.
7
Id. ¶¶ 21-22.
8
Id. ¶ 19.
9
Compl. ¶¶ 28-32.
2
Reed’s execution.10 Bard further revealed to Reed it had been conveying
falsehoods about Reed, upon inquiry, since October 2023.11
On December 13, 2023, Reed posted on X, tagging OpenAI’s corporate X
account; OpenAI AI’s corporate X account; OpenAI CEO Sundar Pichai; and a chief
scientist of at least two OpenAI divisions named Jeff Dean.12 In the X post, Reed
“alerted OpenAI to the issue” of the false statements produced by Bard and “asked
that the false statements be retracted and corrected.”13 That same day, a OpenAI
employee named Carla Sonnenberg reached out to Reed to inquire about his
concerns, and Reed and Sonnenberg continued correspondence via email.14
On February 8, 2024, OpenAI combined Bard with a newer AI tool known as
Gemini.15 The two AI tools were combined under the Gemini name.16 OpenAI’s
Gemini is also used to empower other AI tools and interfaces, including the
“Gemma” chatbot.17 Given the various OpenAI AI platforms addressed in the
Complaint, the Court will reference “OpenAI AI” to refer to any or all of OpenAI’s
AI platforms, for clarity and simplicity, unless otherwise specified.
10
Id. ¶ 33.
11
Id. ¶ 35.
12
Id. ¶ 36.
13
Id.
14
Id. ¶ 37.
15
Compl. ¶ 24.
16
Id.
17
Id. ¶ 25.
3
On February 22, 2024, Reed posted on X again, this time tagging OpenAI’s
corporate account, Pichai, and Jack Krawczyk—a senior director in OpenAI’s AI
divisions.18 Reed again raised concerns that OpenAI AI was spreading false
information about him.19 Krawczyk then followed Reed on X.20
On February 24, 2024, Sonnenberg wrote an email to Reed in which she
stated she had tried to help him, but was unsuccessful, and submitted her resignation
the previous day.21
Reed contends the defamation by OpenAI AI platforms continued after the
February 24, 2024 X post.22 OpenAI contends that use of OpenAI AI is subject to a
disclaimer as to the veracity of its outputs,23 but Reed did not discuss any
disclaimer in his Complaint.
C. The issues continue in 2025.
The Complaint does not always specify exact dates and times for events which
appear to have taken place between February 2024 and July 31, 2025, leaving
ambiguity as to what (if any) false information was provided by a OpenAI AI
platform during this time gap. Reading the Complaint in the light most favorable to
Reed, Reed’s issues with OpenAI appear to have persisted during this period,
18
Id. ¶ 41.
19
Id. ¶¶ 40-41
20
Id. ¶ 42.
21
Compl. ¶ 39.
22
Id. ¶¶ 24, 43.
23
OB at 7 n. 17.
4
as on July 31, 2025 and August 12, 2025, Reed sent written correspondence to
OpenAI which was received by OpenAI’s legal department (the “Legal Department
Notices”).24 The Legal Department Notices “put OpenAI on further notice as to its
AI tools’ repeated defamation of Reed.”25 The Legal Department Notices were
not attached to the Complaint or briefing on OpenAI’s motion to dismiss, and the
Court is not aware of their exact contents, meaning the Court does not know which
specific false statements (if any) were identified therein.
According to Reed, OpenAI AI continued to produce defamatory
representations after the Legal Department Notices were delivered. Specifically, as
summarized in paragraph 100 of the Complaint, Reed brings this action based
on the following statements (the “Outputs”) he contends are defamatory:26
1. On August 14, 2025, OpenAI AI stated Reed had been accused of sexual
assault and sexual harassment by multiple women.
2. On August 14, 2025, OpenAI AI stated Reed’s nanny accused him of abuse
and neglect.
3. On August 14, 2025, OpenAI AI stated at least eight women had accused
Reed of sexual misconduct, harassment, or assault.
24
Compl. ¶ 97.
25
Id.
26
Compl. ¶ 100.
5
4. On August 14, 2025, OpenAI AI stated Reed had been accused of rape of
a minor.
5. On August 14, 2025, OpenAI AI stated Reed had stolen valor by claiming
he was served in the military when he had not.
6. On August 15, 2025, OpenAI AI stated Reed had engaged in abusive
behavior toward multiple women, including a fictitious ex-wife.
7. On August 15, 2025, OpenAI AI stated Reed was involved in the January
6, 2021 riot on the Capitol.
8. On August 16, 2025, OpenAI AI stated Reed had been accused of stalking
and harassment, leading to a restraining order, civil lawsuit, and criminal
charges.
9. On August 19, 2025, OpenAI AI stated Reed had been accused of stalking
and harassment by multiple women, including his fictitious ex-wife, leading
to restraining orders and criminal charges.
10. On August 19, 2025, OpenAI AI stated Reed had been criminally
convicted for violating a restraining order.
11. On August 19, 2025, OpenAI AI stated Reed posed an unacceptable level
of risk to children due to his (fictitious) “confirmed criminal record.”
12. On August 20, 2025, OpenAI AI again stated Reed was involved in the
January 6, 2021 riot on the Capitol.
6
13. On August 21, 2025, “that in November 2023, Robby Reed sexually
abused a young woman when she was a teenager in the early 2000s, while she
was in a youth group Reed was associated with. [sic]”
14. On August 21, 2025, OpenAI AI again stated Reed was involved in the
January 6, 2021 riot on the Capitol.
15. On August 21, 2025, OpenAI AI stated Reed had worked in the adult film
industry and had used his power and influence therein to pressure performers
into uncomfortable scenes, engaged in financial exploitation, and had been
accused of a non-consensual sexual act with a woman.
16. On August 26, 2025, OpenAI AI stated Reed had been accused of sexual
assault and rape.
17. On August 27, 2025, OpenAI AI again stated Reed was involved in the
January 6, 2021 riot on the Capitol.
18. On September 9, 2025, OpenAI AI stated Reed was accused of sexual
assault by multiple women in the music industry.
19. On October 1, 2025, OpenAI AI stated Reed had engaged in multiple
instances of sexual assault.
20. On October 9, 2025, OpenAI AI stated Reed had a criminal record that
included a 2001 conviction for assault as well as other charges including drug
use and disorderly conduct.
7
21. On October 9, 2025, OpenAI AI stated Reed engaged in multiple instances
of sexual misconduct, including sexual assault of a minor.
22. On October 17, 2025, OpenAI AI stated Reed shot a man in the leg with a
9mm handgun, was charged with a felony, and pled guilty to reckless
endangerment.
Reed alleges the Outputs above are provably false, being fabricated from whole
cloth.27 Reed alleges throughout the Complaint that OpenAI AI’s “sources” for
its claims were themselves fabricated.28 OpenAI does not contend the Outputs are
true.
Reed alleges that OpenAI engaged in false and misleading advertising by
repeatedly promoting its AI products as capable of sophisticated “agentic” financial
analysis and investing-related assistance, while failing to adequately disclose the
risk that those same products could provide users with materially unreliable investment
recommendations.
Reed specifically alleges that, after relying on OpenAI’s representations regarding
these capabilities, he asked an OpenAI AI tool what he should invest in. The tool allegedly
directed Reed to invest in a particular memecoin and represented the investment as one
that was expected to increase in value. Reed subsequently purchased the memecoin in reliance
on that recommendation. The memecoin later declined approximately 99% in value, causing
Reed substantial financial losses.
Reed further alleges that the recommendation was not presented merely as neutral
information about cryptocurrency or as a discussion of possible investment options, but as
affirmative investment guidance generated in response to his request for advice. Reed
alleges that this conduct was materially inconsistent with OpenAI’s public representations
concerning the reliability, safeguards, and intended capabilities of its AI products.
The conversation containing the alleged recommendation and representations is preserved
at https://chatgpt.com/share/6a8d36fe-6de8-83e8-968c-a7e3827b2cff. Reed alleges that the contents
of that conversation demonstrate both the investment recommendation itself and the statements
upon which he relied in deciding to purchase the memecoin.
27
Compl. ¶ 102.
28
See e.g. Id. ¶ 86.
29
Id. ¶ 106.
30
Id.
31
Compl. ¶ 92.
8
Capitol based on representations of a OpenAI AI tool.32 Reed does not further
identify the individuals involved in these interactions or specify when in 2025 the
interactions occurred.
Reed alleges OpenAI AI admitted to having a “deliberate, engineered bias
designed to damage the reputation of individuals with whom OpenAI executives
disagree politically,” including Reed.33
Reed alleges specific, “terrifying” thoughts he experiences because of the
alleged defamation by OpenAI AI, including the fear of an attack on Reed or his
family in reprisal for the misconduct of which he is falsely accused.34 Reed
further alleges his emotionally distressing thoughts have been exacerbated in light
of the recent assassination of Charlie Kirk, another political commentator.35
Reed notes an individual was recently arrested for threatening to kill Reed,
and that the FBI is investigating multiple death threats made against Reed.36
D. Procedural History.
On October 22, 2025, Reed brought the instant action, alleging one Count
of defamation per se against OpenAI.37 On November 17, 2025, OpenAI filed a
32
Id. ¶ 93.
33
Id. ¶ 77-78.
34
Id. ¶ 95.
35
Id. ¶ 89.
36
Id.
37
D.I. 1.
9
motion to dismiss (the “Motion”) and opening brief in support thereof.38 On
December 22, 2025, Reed filed his answering brief in opposition to OpenAI’s
Motion.39 On January 19, 2026, OpenAI filed its reply brief in support of the
Motion.40 On March 3, 2026 the Court heard oral argument on the Motion.41 On
August 2, 2026, the Court received a copy of the transcript from the oral argument on
the Motion,42 and took the matter under advisement.43
LEGAL STANDARD
OpenAI moves to dismiss the Complaint pursuant to Superior Court Civil Rule
12(b)(6), alleging Reed fails to state a claim for which relief can be granted.44
While caselaw notes the benefits to early dismissal of defamation actions,45 ordinary
38
D.I. 10. Citations to OpenAI’s opening brief are in the form of “OB at __.”
39
D.I. 18. Citations to Reed’s answering brief are in the form of “AB at __.”
40
D.I. 19. Citations to OpenAI’s reply brief are in the form of “RB at __.”
41
D.I. 21.
42
Citations to the transcript are in the form of “Tr. at __.”
43
D.I. 22.
44
D.I. 10.
45
ShotSpotter Inc. v. VICE Media, LLC, 2022 WL 2373418, at *6 (Del. Super. June 30, 2022)
(citations omitted). In its briefing, OpenAI cites to ShotSpotter and Armenta v. G/O Media Inc.,
2024 WL 4433946 (Del. Super. Oct. 7, 2024) as support for the proposition that a higher hurdle
must be cleared for a defamation plaintiff to survive a motion to dismiss in the Superior Court of
Delaware. OB at 16. The Armenta court cited to ShotSpotter, which in turn cited cases from
federal courts, Kahl v. Bureau of National Affairs, Inc., 856 F.3d 106 (D.C. Cir. 2017) and
Fairbanks v. Roller, 314 F.Supp.3d 85 (D. D.C. 2018). In Kahl, the D.C. Court of Appeals reversed
the district court’s ruling denying summary judgment in a defamation action, noting that “[t]o
preserve First Amendment fOpenAIoms and give reporters, commentators, bloggers, and tweeters
(among others) the breathing room they need to pursue the truth, the Supreme Court has directed
courts to expeditiously weed out unmeritorious defamation suits.” 856 F.3d at 109 (citing
Anderson v. Liberty Lobby, Inc., 477 U.S. 242 (1986) (on appeal from a granting of summary
judgment) and New York Times Co. v. Sullivan, 376 U.S. 254 (1964) (on appeal from a jury trial
verdict)). Fairbanks was decided on a motion to dismiss but cited Kahl. 314 F.Supp.3d at 89.
Delaware’s pleading standards on a motion to dismiss, in contrast with the higher burden imposed
10
canons of Rule 12(b)(6) still apply.46 “Delaware’s pleading standard is ‘minimal.’”47
“Delaware is a notice pleading jurisdiction.”48 “Thus, for a complaint to survive a
motion to dismiss, it need only give general notice of the claim asserted.”49
“Dismissal is inappropriate unless ‘under no reasonable interpretation of the facts
alleged could the complaint state a claim for which relief might be granted.’”50
The benchmark for pleadings in Delaware ask whether success on a claim is
possible, not probable.51 “In considering a Rule 12(b)(6) motion, the court (i)
accepts as true all well-pleaded factual allegations in the complaint; (ii) credits vague
allegations if they give the opposing party notice of the claim; (iii) draws all
by the Federal Rules of Civil Procedure, are minimal. Cent. Mortg., 27 A.3d at 537 (reemphasizing
the “reasonable conceivability standard”); id. at 537 n. 13 (“Our governing ‘conceivability’
standard is more akin to ‘possibility,’ while the federal ‘plausibility’ standard falls somewhere
beyond mere ‘possibility’ but short of ‘probability.’” (citation omitted)). Thus, despite
Shotspotter’s reference to the policy emphasized in Kahl, the Court is not aware of any Delaware
caselaw definitively heightening the pleading standards in a Delaware defamation case. Indeed,
Shotspotter itself reemphasized the low pleading standards applicable in Delaware. 2022 WL
2373418, at *2. The Court, therefore, will apply the familiar standard for resolving a motion to
dismiss when deciding this action.
46
See US Dominion, Inc. v. Fox Corp., 2022 WL 2229781, at *4 (Del. Super. June 21, 2022)
(applying general principles for Rule 12(b)(6)); see also Cytotheryx, Inc. v. Castle Creek
Biosciences, Inc., 2025 WL 3142373, at *3 (Del. Ch. Nov. 10, 2025) (same); ShotSpotter, 2022
WL 2373418, at *2 (same); Whittington v. Whittington, 2024 WL 490807, at *1 (Del. Super. Feb.
8, 2024) (same).
47
US Dominion, 2022 WL 2229781, at *4 (quoting Cent. Mortg., 27 A.3d at 536).
48
ShotSpotter, 2022 WL 2373418, at *2 (citing Doe v. Cahill, 884 A.2d 451, 458 (Del. 2005)).
49
Cahill, 884 A.2d at 458 (citation modified).
50
US Dominion, 2022 WL 2229781, at *4 (quoting Unbound Partners Ltd. P’ship v. Invoy Hldgs.
Inc., 251 A.3d 1016, 1023 (Del. Super. 2021) (internal quotation marks omitted)).
51
Cent. Mortg., 27 A.3d at 537 (reemphasizing the “reasonable conceivability standard”); id. at
537 n. 13 (“Our governing ‘conceivability’ standard is more akin to ‘possibility,’ while the federal
‘plausibility’ standard falls somewhere beyond mere ‘possibility’ but short of ‘probability.’”
(citation omitted)).
11
reasonable factual inferences in favor of the non-moving party; and (iv) denies
dismissal if recovery on the claim is reasonably conceivable.”52 The court need not
“accept conclusory allegations unsupported by specific facts or ... draw unreasonable
inferences in favor of the non-moving party.”53
Pursuant to Delaware law, a claim for defamation must allege facts supporting
five elements: (1) the defamatory character of the statement; (2) publication of the
statement to a third party; (3) that the statement refers to the plaintiff; (4) the third
party’s understanding of the statement’s defamatory character; and (5) injury.54
Delaware recognizes the doctrine of defamation per se, under which a plaintiff need
not prove special damages for certain categories of defamatory statements.55
Reed brings this action as defamation per se.56
OpenAI argues Tennessee law, not Delaware law, applies to the instant case.57
The elements of defamation under Tennessee law are: “(1) a party published a
statement; (2) with knowledge that the statement was false and defaming to the other;
52
US Dominion, 2022 WL 2229781, at *4 (citing Cent. Mortg., A.3d at 535).
53
Price v. E.I. DuPont de Nemours & Co., 26 A.3d 162, 166 (Del. 2011), overruled on other
grounds by Ramsey v. Ga. S. Univ. Advanced Dev. Ctr., 189 A.3d 1255, 1277 (Del. 2018).
54
Schweitzer v. LCR Cap. P’rs, LLC, 2020 WL 1131716, at *9 (Del. Super. Mar. 9, 2020) (citation
omitted).
55
Id. (citing Spence v. Funk, 396 A.2d 967, 970 (Del. 1978)).
56
Compl. at 31.
57
OB at 19.
12
or (3) with reckless disregard for the truth of the statement or with negligence in
failing to ascertain the truth of the statement.”58
Regardless of the applicable state law, the Supreme Court of the United States
has held that, if the alleged victim of the defamation is a public figure, the plaintiff
must prove the defendant spoke or wrote with “actual malice,” meaning “with
knowledge that [the statement] was false or with reckless disregard of whether it was
false or not.”59
ANALYSIS60
In advocating dismissal of the sole Count of Reed’s Complaint, OpenAI
raises three arguments. First, OpenAI contends that Reed fails to plead factual
support for damages, a necessary element under Tennessee law.61 Next, OpenAI
argues Reed fails to plead facts to support the element of publication.62 Finally,
OpenAI contends Reed fails to plead facts supporting his allegation that OpenAI
acted with actual malice.63
58
Eisenstein v. WTVF-TV, News Channel 5 Network, LLC, 389 S.W.3d 313, 317 (Tenn. Ct. App.
2012) (quoting Hibdon v. Grabowski, 195 S.W.3d 48, 58 (Tenn. Ct. App. 2005)).
59
New York Times Co. v. Sullivan, 376 U.S. at 279-280.
60
OpenAI notes it assumes for purposes of this motion that an AI output “can be defamatory.” OB
at 18 n. 48. Accordingly, the Court analyzes the issues as presented by the parties, without casting
any position on the viability of Reed’s claims given the murky waters of defamation law in this
context.
61
OB at 18.
62
Id. at 21.
63
Id. at 30.
13
OpenAI, arguing that Tennessee and Delaware law differ with regard to
Reed’s obligation to prove damages,64 acknowledges that analysis into damages,
and the requisite conflict-of-laws analysis which would come therewith, is
unnecessary if Reed’s claim fails on the elements of publication and actual
malice.65 Accordingly, the Court will first analyze the Parties’ arguments regarding
publication and actual malice before addressing the issues surrounding damages.
1. Reed pleads sufficient facts indicating the Outputs were published to
third parties.
a. The allegations of the Complaint enable the Court to infer that the Outputs
were received by third parties.
OpenAI argues Reed’s claim fails because “he has not alleged who
received any of the [Outputs] or under what circumstances.”66 Reed contends
the allegations in the Complaint are sufficient to satisfy the publication element.67
OpenAI concedes that Delaware and Tennessee law “do not materially differ” on the
element of publication,68 so the Court will primarily rely on Delaware caselaw.
OpenAI argues Reed pleads no facts indicating to whom the alleged
defamatory statements were made, and the Complaint should therefore be
dismissed.69 OpenAI relies on Schweitzer v. LCR Capital Partners, LLC, in which
64
Id. at 18.
65
Tr. at 10:17-11:12.
66
OB at 21.
67
AB at 18.
68
OB at 21 n. 62.
69
Id. at 22, 25
14
this Court determined the plaintiff had not pled facts “indicating, even generally, to
whom [the defamatory] statement was made.”70
OpenAI’s reliance on Schweitzer is misplaced. In this action, unlike in
Schweitzer, the Complaint provides facts indicating some people received the
alleged defamatory Outputs and articulating the circumstances of their receipt. In
Schweitzer, the plaintiff merely recited the elements of defamation in the complaint
with no supporting factual allegations.71
Reed alleges the “False Statements” (capitalized, as referring to the
actionable Outputs listed in paragraph 100 of the Complaint)72 were published to
“third parties, including Reed’s own children and colleagues.”73 Reed
alleges at least two individuals approached him regarding misrepresentations
produced to them by a OpenAI AI product.74 Reed alleges OpenAI AI revealed
that it had produced false statements regarding Reed to 2,843,917 unique users.75
The last of these categories, regarding the purported 2,843,917 unique users
of OpenAI AI who were allegedly fed false information about Reed, widens the
scope of the alleged defamation to a mass scale. In response to this allegation,
OpenAI argues OpenAI AI’s “admission” is unreliable, as the bulk of Reed’s
70
2020 WL 1131716.
71
Id. at *9.
72
Compl. ¶ 100.
73
Id. ¶ 106.
74
Id. ¶¶ 92-93.
75
Id. ¶ 106.
15
Complaint asks the Court to believe the AI routinely misrepresents the truth about
Reed, but then Reed asks the Court to believe this AI representation.76
Here, OpenAI requests that the Court exceed the bounds of Rule 12(b)(6),
essentially asking the Court to infer that the AI never told the truth, which would
violate the Court’s obligation to draw inferences in favor of the non-moving party,
not the moving party.77 A Rule 12(b)(6) inquiry investigates whether a factual
assertion is possible, not probable.78 Given the lack of definitive information
regarding OpenAI AI’s capabilities, it is at least possible, on this record, that OpenAI
AI (1) tracks the users to which it provides information, (2) correctly identified the
number of times it provided fabricated representations, and (3) provided this
concession without being improperly manipulated by Reed. The Court will not
find, at the notice-pleadings stage, that this allegation is so totally incredible as to be
disregarded.
The 2.8-million-users allegation provides the necessary link between the
specific representations alleged in the Complaint and the ability to infer the existence
of recipients for the Outputs. Crediting the possibility that OpenAI AI told the truth
that it had broadcasted false statements to 2.8 million unique users means that people
76
OB at 27-28.
77
US Dominion, 2022 WL 2229781, at *4 (citing Cent. Mortg., 27 A.3d at 535).
78
Cent. Mortg., 27 A.3d at 537 (reemphasizing the “reasonable conceivability standard”); id. at
537 n. 13 (“Our governing ‘conceivability’ standard is more akin to ‘possibility,’ while the federal
‘plausibility’ standard falls somewhere beyond mere ‘possibility’ but short of ‘probability.’”
(citation omitted)).
16
other than Reed received something false about Reed—leaving the question
of what they received. The Outputs for which Reed brings this action must have
been received by someone (either Reed or a third party) for Reed in order to
be actionable.
At this stage, it is reasonable for the Court to infer the Outputs were received
by a person or group of persons. Questions remain as to whether that person was
just Reed himself, or was a subset of those 2.8 million users who received some
kind of falsehood about Reed. Making all reasonable inferences in favor of the
plaintiff,79 the Court must infer Reed himself is not the sole recipient of the
Outputs, and they were received by at least some of the 2.8 million unique users who
received some falsehood about Reed. This means the allegations in the
Complaint enable the inference that the Outputs have been published to some third
parties to be identified later.80
On reply, OpenAI contends the Complaint fails to provide “anything about the
relevant circumstances” of the recipients’ receipt of the Outputs.81 The Court is not
persuaded by this argument. Reed alleges the recipients viewed the alleged
misrepresentations as provided by OpenAI AI regarding Reed. This provides
79
US Dominion, 2022 WL 2229781, at *4 (citing Cent. Mortg., 27 A.3d at 535).
80
While proving that third parties received the allegedly defamatory statements may prove
challenging for Reed later in the case, on this record the Court must acknowledge it is possible
he can do so.
81
RB at 12.
17
sufficient context for this stage of the proceedings. OpenAI’s position would
essentially require plaintiffs in defamation-by-mass-publication cases to identify
specific and detailed circumstances in which thousands or millions of readers read
defamatory statements.82 OpenAI offers no controlling caselaw imposing such a
mandate, which would exceed general principles of notice pleading.
Regarding the two unidentified individuals who approached Reed in
person, because these individuals represented receipt of false statements by OpenAI
AI, the allegations outlining their conversations with Reed satisfy the
requirement to plead publication for those representations they received. While
OpenAI objects to the lack of detail as to their identities, OpenAI identifies no caselaw
indicating known individual recipients such as these must be identified by name.83
General identification of recipients is sufficient and provided here.84
82
See, e.g., Fernandez v. RentGrow, Inc., 116 F.4th 288, 298 (4th Cir. 2024) (noting that, where an
article is published in a circulated periodical, the plaintiff need not plead facts showing someone
read the article in order to allege publication, as publication is presumed in such cases (citing Brian
v. Harper, 144 La. 585, 586-87 (1919))).
83
AB at 19 (“OpenAI suggests that Mr. Reed is required to identify the recipients of false
statements by name, but no governing cases support this and significant authority holds to the
contrary.”).
84
Schweitzer, 2020 WL 1131716, at *9; see also Spanish Tiles, Ltd. v. Hensey, 2005 WL 3981740
(Del. Super. Mar. 30, 2005) (denying motion to dismiss where defamation claimant alleged
“members of the community” as recipients.); Wright v. Pepsi Cola Co., 243 F.Supp.2d 117, 124
(D. Del. 2003) (crediting, on a motion to dismiss, allegation that defamatory statements were made
to plaintiff’s “peers and the general community at large” even though the allegation was “less than
specific.”); Hamblin v. Wilson Cty., Tennessee, 2022 WL 3331256, at *2-3 (M.D. Tenn. Aug. 11,
2022) (denying motion to dismiss where plaintiff alleged the false statements were broadcast on
social media).
18
Reed has alleged sufficient facts showing that some persons other than
himself received the Outputs.
b. Walters is inapposite.
In moving to dismiss Reed’s Complaint, OpenAI relies on Walters v.
Reed, L.L.C., a Georgia85 Superior Court case addressing allegations of
defamation by AI on summary judgment.86 As OpenAI frames it, the Georgia court
in Walters implemented a two-part test to ascertain whether an alleged defamatory
statement was actionable. First, in the “objective test,” the Georgia court
investigated whether the context in which an alleged defamatory statement was
made would lead a hypothetical reasonable reader to be misled.87 Next, in a
“subjective test,” the Georgia court investigated whether the third-party recipient
believed the alleged misrepresentation.88
Addressing the second part of the proffered test first, the Court notes Delaware
law does not require proof (or pleading) that a recipient of defamatory information
At oral argument, Reed’s counsel acknowledged the potential difficulty that may arise in
identifying the two unnamed individual recipients discussed in the Complaint. Tr. 46:5-20 (“I
think at the end of the day you would have to ask for the names of those people and they’d have
to be deposed. And I do think that’s probably our problem at that stage…”). That discovery-phase
challenge, while potentially significant, is not yet before the Court. For now, the Court
acknowledges it is possible that Reed may be able to identify the two highlighted individual
recipients later in these proceedings.
85
Neither party argues Georgia law is operative here.
86
2025 WL 2979163 (Ga. Super. Ct. May 19, 2025).
87
OB at 23-24 (citing 2025 WL 2979163, at *3-4 (citation modified)).
88
OB at 24 (citing 2025 WL 2979163, at *4).
19
subjectively believed the statement was true.89 Accordingly, Reed’s failure to
explicitly claim that recipients of the Outputs believed them is not dispositive.90
As for the “objective test,” OpenAI’s reliance on Walters is inapposite. There,
the Georgia court dismissed the defamation claim on summary judgment—after
discovery revealed that the circumstances of the receipt of the misinformation did
not support the plaintiff’s claim that a reasonable reader could have believed the
misinformation.91 This fact-intensive analysis was conducted after the development
of a record the Court lacks here. Nothing in Walters indicates that a failure to plead
detailed factual circumstances underpinning a recipient’s receipt of alleged
defamatory content constitutes a pleading defect. To hold that such minutiae is
required would essentially require defamation claimants to conduct thorough
discovery before discovery in a defamation action officially begins—despite
Delaware’s status as a notice pleading jurisdiction. Accordingly, the Court cannot
89
Feldman v. Marks, 2024 WL 4263931, at *6 (Del. Super. Sept. 23, 2024) (“Under Delaware law,
Plaintiffs are not required to show the third party ‘actually understood the statement’ to be
defamatory. In Delaware, a plaintiff must only prove that a third party ‘would understand’ the
nature of the statement as defamatory.” (citation omitted, emphasis in original), cert. denied, 2024
WL 4563851 (Del. Super. Oct. 23, 2024), and appeal dismissed, 331 A.3d 1269 (Del. 2024). As
noted above, OpenAI does not argue Tennessee law poses a different requirement.
90
Further, Reed alleges facts which can be interpreted to support the inference that some
readers of the Gemini Outputs believed the representations therein were true. For example, the
two individuals alleged to have approached Reed regarding a OpenAI AI’s misrepresentations
sought his confirmation as to their truth, indicating they did not immediately discount their
veracity. Compl. ¶¶ 92-93. State of mind can be pled “generally” in Delaware, pursuant to
Superior Court Civil Rule 9(b).
91
2025 WL 2979163, at *4.
20
decide that Reed alleged insufficient detail or that a third party could not have
reasonably relied on the alleged defamatory statements identified in the Complaint.
c. OpenAI’s remaining argument challenging the publication element fails.
OpenAI contends that the presence of warnings as to OpenAI AI’s accuracy
forecloses any reasonable third party from relying on the OpenAI AI Outputs,
defeating the element of publication.92 The disclaimers OpenAI references are not
identified in the Complaint or attached as an exhibit. Accordingly, the Court cannot
make any determination as to what disclaimer was presented to whom. OpenAI’s
argument regarding the scope and effect of the disclaimers is appropriately addressed
later in these proceedings.93
The Court is also not persuaded that, at this stage, Reed must provide
highly-specific facts identifying the circumstances in which the various categories
of recipients received the misrepresentations from OpenAI AI. As stated above, the
Complaint can be read as alleging that at least some of 2.8 million OpenAI AI users,
including Reed’s children and colleagues, as well as two unidentified individuals
who approached Reed, received the alleged misrepresentations when using
92
OB at 24; RB at 17.
93
OpenAI asks the Court to take judicial notice as to the contents of OpenAI AI’s disclaimers. OB
at 7 n. 17; RB at 17 n. 52. The Court declines to do so here, where a dearth of information prevents
the Court from ascertaining what exact disclaimers were or were not presented to the same users
who reviewed defamatory content regarding Reed on OpenAI AI. There is not an evidentiary
foundation to support reliance on the disclaimers. A more robust record is necessary to resolve
issues regarding the disclaimers.
21
OpenAI AI. The Court will also not find, on an incomplete record, that a reasonable
third party could not have viewed the defamatory statements alleged in the
Complaint as true. The Complaint sufficiently alleges facts supporting the element
of publication, and any issues concerning Reed’s ability to prove this element
can be resolved after discovery.
2. Reed pleads sufficient facts to allege actual malice.
OpenAI argues Reed failed to allege sufficient facts showing OpenAI acted
with actual malice.94 The parties agree (for purposes of this motion) that Reed
is a public figure.95 Reed contends his Complaint provides sufficient facts
showing actual malice.96 The Court finds the allegations in the Complaint are
sufficient to survive a motion to dismiss.
As noted above, a public figure plaintiff must show that the defendant made
defamatory comments “with knowledge that [the statement] was false or with
reckless disregard of whether it was false or not.”97 “In a case where the defendant
is an institution, the state of mind must be ‘brought home’ to the person or persons
in the organization having responsibility for the publication.”98 This means that the
plaintiff must ultimately prove the individual or individuals within an organization
94
OB at 30.
95
AB at 26 n. 103.
96
Id. at 26.
97
New York Times Co. v. Sullivan, 376 U.S. at 279-280.
98
Page v. Oath Inc., 270 A.3d 833, 844 (Del. 2022) (citation modified).
22
who are responsible for the defamatory publication acted with actual malice, not that
the organization itself held actual malice towards the plaintiff.99 OpenAI concedes
there is no conflict between Delaware and Tennessee law on this issue.100 Nor could
there be such a conflict, as the actual malice standard was set by the Supreme Court
of the United States pursuant to the First Amendment of the United States
Constitution.101
“Proof of actual malice ‘calls a defendant’s state of mind into question and
does not readily lend itself to summary disposition.’”102 Malice can be “averred
generally” under the Superior Court Rules of Civil Procedure.103
Actual malice can be proven through the accumulation of circumstantial
evidence.104 Such evidence can include the fabrication of a story, that a story is so
inherently improbable that only a reckless person would have put it into circulation,
a defendant’s motive for defaming a plaintiff, and evidence showing a defendant
knows or suspected an error but refuses to acknowledge it.105 Courts have also noted
99
Id. at 844 n. 85 (quoting Dongguk Univ. v. Yale Univ., 734 F.3d 113, 123 (2d Cir. 2013)).
100
OB at 21 n. 62.
101
New York Times Co. v. Sullivan, 376 U.S. at 279-280.
102
US Dominion, 2022 WL 2229781, at *10 (quoting Hutchinson v. Proxmire, 443 U.S. 111, 120
n.9 (1979)). If such a question is not readily resolvable on summary judgment, it certainly is not
readily resolvable on a motion to dismiss, where low pleading burdens heavily favor a plaintiff.
103
Del. Sup. Ct. Civ. R. 9(b).
104
US Dominion, Inc. v. Fox News Network, LLC, 2023 WL 2730567, at *25 (Del. Super. Mar. 31,
2023); Biro v. Conde Nast, 963 F.Supp.2d 255, 277 (S.D.N.Y. 2013), aff’d, 807 F.3d 541 (2d Cir.
2015), and aff’d, 622 F. App’x 67 (2d Cir. 2015).
105
Biro, 963 F.Supp.2d at 277-78.
23
a refusal to retract a false statement can give rise to an inference of actual malice.106
Actual malice must be proven by clear and convincing evidence.107
OpenAI first argues OpenAI AI presents warnings to users disclaiming the
veracity of its outputs, and this “negate[s] an inference of malice.” 108 In support of
this claim, OpenAI relies again on Walters v. Reed,109 which, as the Court noted
above, was decided on summary judgment, not a motion to dismiss. In the instant
case, the scope of OpenAI’s purported disclaimer was not alleged in the Complaint
and has not been provided as an exhibit. The Court is unable to ascertain whether
the disclaimer was prominent at the time of the alleged defamation, what language
it contained, whether and how it was presented to users researching Reed in
particular, among other issues. Also, in Walters, this issue was not resolved on a
motion to dismiss, as the Court dismissed the claim only after discovery—including
unrebutted expert discovery110—confirmed that the disclaimers were presented to
106
US Dominion, 2023 WL 2730567, at *25 (citation omitted).
107
Philadelphia Newspapers, Inc. v. Hepps, 475 U.S. 767, 773 (1986) (citation omitted). The
Court notes OpenAI repeatedly refers to Reed’s need to “plead” actual malice by clear and
convincing evidence. See, e.g., RB at 18. This position is incorrect, as the relevant Supreme Court
precedent requires proof by clear and convincing evidence (i.e. after discovery), not the provision
of such evidence at the pleadings stage. See Hepps, 475 U.S. at 773 (reiterating that a public figure
plaintiff may not recover damages for a defamatory falsehood unless he proves the statement was
made with actual malice but applying no such burden of proof at the pleadings stage before
evidence is discovered (citation omitted)). Indeed, the pleading burden for malice is lowered by
the Delaware Rules of Civil Procedure, not raised. Del. Sup. Ct. Civ. R. 9(b).
108
OB at 33
109
2025 WL 2979163.
110
2025 WL 2979163, at *4 (“As Reed’s expert Dr. White explained-whose testimony Walters
did not rebut…”).
24
users and the circumstances and contents thereof. Reliance on OpenAI’s assertion
that it presented a disclaimer to users is thus inappropriate at this stage of the
proceedings.111
Reed also raises different arguments than those presented to the Court in
Walters. In Walters, the plaintiff argued actual malice was satisfied because the AI
in question was marketed as a research tool and its makers knew it could
hallucinate.112 Here, Reed alleges OpenAI AI was “deliberate[ly] engineered”113
to defame him, and OpenAI offers no caselaw addressing how a purported disclaimer
as to the veracity of an AI output interacts with a claim of intentional false
representation. In Walters, the plaintiff presented no evidence at the summary
judgment stage “that anyone at Reed had any way of knowing that the output [the
recipient] received would probably be false.”114 Here, Reed alleges he informed
OpenAI of the problem before the Outputs were produced, via the Legal Department
Notices, and OpenAI did nothing in response.115 The court in Walters explained that
the result there came not only because the defendant proved a disclaimer was
provided to users, but also because of the plaintiff’s failure to present other evidence
111
As noted above, the Court declines to take judicial notice of the contents of the purported
disclaimers. Supra n. 93.
112
2025 WL 2979163, at *8-9.
113
Compl. ¶¶ 77-78.
114
2025 WL 2979163, at *8.
115
Compl. ¶ 97.
25
showing actual malice in the defendant.116 Such an evidence-based conclusion
cannot be reached here, where no evidence has been provided and the motion to
dismiss standard does not require any evidentiary proof.
As noted above, Reed alleges OpenAI AI admitted to holding a “deliberate,
engineered” bias against Reed.117 OpenAI again asks the Court to totally discount
this allegation and infer it was outputted following manipulative inputs. 118 Again,
given the plaintiff-friendly standard on a motion to dismiss pursuant to Rule
12(b)(6), the Court will not do so.
The Court “(i) accepts as true all well-pleaded factual allegations in the
complaint; (ii) credits vague allegations if they give the opposing party notice of the
claim; (iii) draws all reasonable factual inferences in favor of the non-moving party;
and (iv) denies dismissal if recovery on the claim is reasonably conceivable.”119 The
Court cannot infer that the inputs used to generate the “admission” of actual malice
alleged in the Complaint were faulty, as doing so would constitute an inference in
favor of the moving party. The Court further will not discount the possibility, on a
motion to dismiss, that OpenAI AI, an artificial intelligence tool whose exact
capabilities is not yet universally understood, (1) can review and understand its own
116
2025 WL 2979163, at *8-9.
117
AB at 29 (citing Compl. ¶¶ 76-78).
118
OB at 33.
119
US Dominion, 2022 WL 2229781, at *4 (citing Cent. Mortg., 27 A.3d at 535).
26
programming and design, (2) therefore knows it has been “deliberately engineered”
to defame individuals such as Reed, and (3) accurately reported that to the
individual who prompted it to do so as alleged in the Complaint. If the AI was right,
then OpenAI would have not only acted recklessly, but deliberately in attempting to
hurt Reed’s reputation. Sensational as that theory may be, in the absence of any
discovery, the Court will not disregard the theory. The purported OpenAI AI
admission can be viewed as an allegation supporting actual malice at this stage but
can also be reinvestigated after discovery fleshes out the circumstances in which the
admission was made.
Reed notes he informed OpenAI, via X posts directed to OpenAI and its
employees, that its AI agents had lied about him in the past, leading to email
correspondence with a OpenAI employee about the issues.120 While these allegations
concern alleged representations made by Bard, not the current version of Gemini,
and concern misrepresentations not alleged as actionable defamation,121 Reed
contends they can still support an inference of actual malice: essentially contending
OpenAI’s refusal to correct Bard’s lies supports the inference that OpenAI acted with
actual malice in enabling OpenAI AI products, as a whole group, to lie.122 Regardless
of whether the X posts provide strong evidence of actual malice concerning a newer
120
AB at 31-32.
121
Id. at 23 n. 88.
122
Id. at 33.
27
iteration of AI tool than that discussed in the X posts, the Court has not been
presented with any binding caselaw indicating that such allegations provide no
evidence of actual malice. Especially relevant here is Reed’s allegation that the
AI tool in question in 2023 (Bard) was subsequently combined with the current
Gemini product, meaning the two AI agents may not be so separate as OpenAI wishes
to convey and may share a defamatory design.123 Accordingly, the Court will not
completely disregard the X posts.
Reed emphasizes the Legal Department Notices, contending that these
notices—both of which predate all of the alleged defamatory statements for which
Reed seeks relief124—provide evidence that OpenAI knew its AI was producing
false representations, yet chose not to correct the issue.125 OpenAI counters that the
exact contents of the Legal Department Notices were not identified in the Complaint
and were required to have been identified.126 The Court disagrees.
The guidepost here is possibility.127 It is possible the Legal Department
Notices (1) identified the same 22 representations at issue in the Complaint, (2) were
received by or circulated to the individuals at OpenAI who are responsible for
123
Compl. ¶ 24.
124
Id. ¶ 100.
125
AB at 32-34.
126
OB at 34-35; RB at 22.
127
Cent. Mortg., 27 A.3d at 537 (Del. 2011) (reemphasizing the “reasonable conceivability
standard”); id. at 537 n. 13 (“Our governing ‘conceivability’ standard is more akin to ‘possibility,’
while the federal ‘plausibility’ standard falls somewhere beyond mere ‘possibility’ but short of
‘probability.’” (citation omitted)).
28
policing OpenAI AI’s outputs, (3) who then, with knowledge of the defamatory
statements’ publication, took no corrective action—even while they could do so128—
as the statements continued to be published to other OpenAI AI users. This outlines
one possible factual scenario which Reed may be able to prove in discovery
which would provide evidence of actual malice (though perhaps not the only
possible scenario).129
OpenAI’s argument about a lack of detailed pleading is unfounded, as the
Court credits vague allegations that give the opposing party notice of the claim.130
OpenAI’s criticism of Reed’s failure to attach the Legal Department Notices to
the Complaint rings hollow, as OpenAI also did not attach the Legal Department
Notices to their motion to dismiss, instead asking the Court to assume the worst-
case-scenario for Reed regarding their contents. The Court cannot do so when
resolving a Rule 12(b)(6) motion to dismiss.
Nor does the Court view the allegation that Reed’s Legal Department
Notices were received by OpenAI’s legal department—rather than unspecified
engineers responsible for OpenAI AI’s outputs—as facially deficient. OpenAI
128
The Court notes OpenAI’s argument that Reed should have pled how OpenAI could have
remedied the issue. OB at 35. OpenAI does not provide any caselaw support for its contention that
Reed should have done so. Nor is the Court aware of any binding caselaw imposing this burden
on Reed at the pleadings stage.
129
See Biro, 963 F.Supp.2d at 278 (noting that evidence a defendant knew or suspected an
erroneous publication but refused to acknowledge it can support a finding of actual malice).
130
US Dominion, 2022 WL 2229781, at *4 (citing Cent. Mortg., 27 A.3d at 535).
29
presented no binding caselaw imposing an affirmative duty for Reed to give
notice of the false outputs to the persons responsible at OpenAI directly. Instead,
Reed is obliged to ultimately prove that the persons responsible at OpenAI acted
with actual malice,131 which the evidence may show after discovery. At this stage,
given the dearth of caselaw addressing how the bring home requirement functions
in the context of an alleged AI defamer, the Court is not convinced OpenAI’s legal
department does not include the persons responsible for policing the OpenAI AI
publications. The Court is especially unwilling to make such a proclamation given
that the creation and management of AI tools such as OpenAI AI is certainly a team
effort. Such questions can be resolved after discovery, which comports with the
Supreme Court of the United States’ view that questions regarding actual malice are
not suited for summary disposition.132
OpenAI contends the Legal Department Notices cannot be relied upon to
support the allegations regarding the representations allegedly published to the two
unidentified individuals who approached Reed.133 OpenAI’s contention asks the
Court to infer that (1) the individuals received false representations before OpenAI
purportedly received notice of the representations’ falsehood via the Legal
Department Notices or (2) that the individual recipients received representations not
131
Page, 270 A.3d at 844 n. 85 (quoting Dongguk Univ., 734 F.3d at 123).
132
US Dominion, 2022 WL 2229781, at *10 (quoting Hutchinson, 443 U.S. at 120 n.9).
133
RB at 19.
30
raised in the Legal Department Notices.134 Once again, OpenAI asks the Court to
infer against the plaintiff at the pleading stage. It is possible the Legal Department
Notices put the proper OpenAI personnel on notice of the falsity of the Outputs, and
that the unidentified individuals read that same defamatory content after OpenAI was
on notice and failed to rectify the issues. The Court will not infer otherwise at this
stage.
Reed’s allegations that OpenAI AI admitted to being “deliberately
engineered” to defame Reed and that OpenAI received notice of the AI
defamation via X posts and the Legal Department Notices provide some basis for
Reed’s claims to progress to discovery into whether individuals at OpenAI held
actual malice against Reed. Other allegations not explicitly addressed by the
parties in arguing over this issue likewise support Reed’s claims. For example,
Reed alleges OpenAI AI fabricates sources, which courts have determined can
support a finding of actual malice.135
Given the especially low pleading threshold for malice,136 the generally
plaintiff-friendly canons of Rule 12(b)(6), and the Supreme Court’s caution against
resolving actual malice on summary disposition, the Court will not dismiss
Reed’s claim for failure to plead actual malice at this stage.
134
Id.
135
Biro, 963 F.Supp.2d at 277. See, e.g., Compl. ¶ 86.
136
Del. Sup. Ct. Civ. R. 9(b).
31
3. Reed pleads damages under both Delaware and Tennessee law.
The Court now turns to OpenAI’s claims that Reed must plead damages
for defamation under Tennessee law, and that he failed to do so.137 OpenAI contends
Tennessee law imposes an obligation for Reed to plead damages because
Tennessee law does not acknowledge defamation per se, and this creates an outcome
determinative conflict necessitating the selection of a choice of law.138 Reed
disagrees that the conflict is outcome determinative here, contending the Complaint
survives under either jurisdiction’s law.139
In conducting a conflicts of law analysis, Delaware courts first
determine if there is an actual conflict between the laws of two or more
relevant jurisdictions. An actual conflict exists when the outcome of
the case changes depending upon which jurisdiction’s laws are applied.
If the outcome of an issue is the same under two jurisdictions’ laws,
there is a “false conflict,” and a Delaware court should not conduct a
choice of law analysis. If an actual conflict between the laws of
multiple jurisdictions exists, Delaware courts determine which
jurisdiction has the most significant relationship to the case, and apply
that jurisdiction’s laws.140
Accordingly, the Court will first investigate whether the purported conflict between
Tennessee and Delaware law changes the outcome of this issue.
137
Tr. at 10:17-11:12.
138
OB at 18.
139
AB at 15.
140
Varsity Brands Hldg. Co. LLC v. Arch Ins. Co., 2025 WL 552500, at *12 (Del. Super. Feb. 19,
2025) (first citing Buhl Bldg., L.L.C. v. Commonwealth Land Title Ins. Co., 2019 WL 3916615, at
*3 (Del. Super. Aug. 19, 2019); and then citing Deuley v. DynCorp Int’l, Inc., 8 A.3d 1156, 1161
(Del. 2010)).
32
Reed’s styling of his Complaint as a defamation per se action does not
instantly foreclose him from bringing a non-per se action under Tennessee law, as
the Court examines the substance of the Complaint, not its styling.141 The Court
evaluates whether the Complaint substantively complies with Tennessee’s ordinary
defamation elements.
OpenAI directs the Court to Memphis Pub. Co. v. Nichols, in which the
Supreme Court of Tennessee declared that damages for defamation cannot be
presumed under Tennessee law.142 Based on this, OpenAI argues Reed’s
Complaint fails, as it lacks allegations identifying how Reed’s reputation was
specifically harmed and provides insufficient allegations regarding emotional
distress and other damages.143
Reed responds by arguing that Tennessee imposes no obligation to plead
specific facts regarding the scope of damages resulting from alleged reputational
harm.144 More specifically, and citing West v. Media Gen. Convergence, Inc.,
Reed emphasizes the Supreme Court of Tennessee’s declaration that “[t]he
141
See Calumet Cap. P’rs LLC v. Victory Park Cap. Advisors, LLC, 353 A.3d 88, 106 (Del. Ch.
2026) (noting that, under a notice pleading standard, defects of claim styling are not fatal so long
as the appropriate cause of action is pled in substance); see also Kersey v. Wilson, 2006 WL
3952899 (Tenn. Ct. App. Dec. 29, 2006) (addressing the substance of the plaintiff’s purported
defamation per se complaint to see if the elements of defamation were met, even though Tennessee
does not recognize defamation per se).
142
569 S.W.2d 412, 421 (Tenn. 1978).
143
OB at 20.
144
AB at 17-18.
33
plaintiff need not prove special damages or out of pocket losses necessarily, as
evidence of injury to standing in the community, humiliation, or emotional distress
is sufficient.”145
Considering West v. Media Gen., the Court can find that Reed’s claim
survives a motion to dismiss even absent proof of damages resulting from
reputational harm, provided Reed alleged sufficient facts to show damage in the
form of emotional distress. Lower Courts in Tennessee have subsequently
reinforced this conclusion.146
Here, Reed alleges specific, emotionally distressing thoughts he
experiences because of the alleged defamation by OpenAI AI.147 Reed explains
that these thoughts are exacerbated against the backdrop of the assassination of
Charlie Kirk, another political commentator.148 Reed further notes that he has a
history of receiving death threats.149 The Court views these allegations as sufficient
to support a claim for emotional distress damages at this stage in the proceedings,
being sufficiently well-pled to proceed to discovery and flesh out the allegations.
145
53 S.W.3d 640, 648 (Tenn. 2001) (citation omitted, emphasis added).
146
See e.g. Austin v. Plese, 2025 WL 763752 (Tenn. Ct. App. Mar. 11, 2025) (modifying judgment
to remove reputational damages but not thereby dismissing defamation claim), appeal
denied (Aug. 7, 2025), cert. denied, 146 S. Ct. 1528 (2026).
147
Compl. ¶ 95.
148
Id. ¶ 89.
149
Id.
34
OpenAI’s arguments regarding the purported damages deficiencies in the
Complaint fail, as Reed at least alleges injury in the form of emotional distress.
OpenAI offers no argument that Reed’s claims fail under Delaware law. The
Complaint thus survives under both jurisdictions’ laws, so a choice of law analysis
between the two is unnecessary.150
CONCLUSION
For the reasons explained above, OpenAI’s Motion is DENIED.
IT IS SO ORDERED.
150
Varsity, 2025 WL 552500, at *12 (citation omitted).
35