Andrew Bridenstine, Associate Member 2025-2026
Intellectual Property and Computer Law Journal
I. Introduction
One morning, you wake up to a slew of messages from friends, family, and coworkers accusing you of posting something completely false on social media. Further investigation indicates that the inflammatory post was generated by artificial intelligence (AI). False information that is produced by generative AI is becoming an increasing issue with the rise of large language models.
This blog discusses specific cases when AI has produced false information regarding individuals, and the possible legal recourse for those people. Part II examines the current landscape of defamation laws and specific instances where defamation has been at the hands of AI. Part III of this blog discusses how lawmakers must update current defamation statutes to add provisions holding AI companies liable for misinformation produced by their products.
II. Background
The World of AI
The use of AI has greatly appreciated, and it is now being used by almost everyone, whether they know it or not. The current state of AI can be broken down into subcategories such as reactive, predictive, and generative.[1] Reactive AI is a tool that responds to specific inputs without learning from past experiences, whereas predictive AI uses historical data to predict future events.[2] Generative AI models, like ChatGPT, generate responses based on data patterns and structures.[3] Although generative AI may seem like the most popular today, reactive and predictive AI models have been used for a long time.[4] This blog focuses primarily on generative AI.
With the growing popularity of these tools, especially generative AI models, more conversation has been directed towards the implications of AI capabilities.[5] One problem that has stemmed from the heightened use is AI hallucinations, when inaccurate content is produced.[6] These occur because generative AI models are trained on internet data that contains both accurate and inaccurate content.[7] As a result, individuals have unsurprisingly fallen victim to generative AI models producing false statements about them.
Current Defamation Landscape
Whether AI-generated statements can satisfy the legal elements of defamation has become a question for the courts. To[CL1] succeed on a defamation claim, a person must prove four things: 1) a false statement purporting to be a fact; 2) publication or communication of that statement to a third person; 3) fault amounting to at least negligence; and 4) there was some harm caused to the reputation of the person.[8] As generative AI use increases rapidly, the risk that false statements will be generated, widely disseminated, and cause reputational harm also increases.
The rise in false statements stems from the increased availability of generative AI tools at the fingertips of nearly everyone and the realism of their AI-generated content, which has made spotting what is true and false more difficult. Over 77% of devices in use contain some form of AI capability, and current projections estimate the growth of AI to be at least 120% year-over-year.[9] Generative AI, such as ChatGPT, produces responses by predicting likely text from patterns learned during training; it does not retrieve or fact-check like a search engine. That is why an answer can read correctly yet be false.
This production of false information has led several public figures to take legal action against major AI companies for false statements that stem from generative AI outputs.[10] Unfortunately, these lawsuits will be difficult due to the confusion surrounding Section 230, and companies claiming protection. However, those protections should not extend to generative AI companies.
Section 230 of the Communications Decency Act was created to shield companies from defamation liability.[11] This means that while Section 230 protects platforms such as Reddit and Facebook from liability for user-generated content, it does not automatically extend to content generated by AI.[12] This problem is not uniquely American. Due to the limited court precedent, other countries have started to adopt frameworks that address the challenges with the rise of AI.[13] The American legal system needs to follow other countries by either creating new statutes and regulations or updating existing ones, like Section 230, to ensure protection from AI defamation.
Instances of AI Defamation and Outcomes
With the growing popularity of generative AI, reports of AI models producing inaccurate outputs and hallucinating biographical data have begun to surface. In April of 2025, the conservative activist Robby Starbuck alleged that Meta AI falsely accused him of participating in the January 6th Capitol riot.[14] Although this case has not yet been decided, Starbuck v. Meta could end up setting a precedent for AI defamation cases by establishing the way American courts will govern generative AI-based defamation.
The Starbuck v. Meta case was not the first to reach U.S. courts; generative AI defamation was also the subject of Walters v. Open AI, LLC. That case was filed in June of 2023, after a reporter incorporated a false story generated by the platform.[15] Walters sued the owner and developer of ChatGPT, OpenAI, for generating the summary that falsely stated he had been accused of embezzlement.[16] In this case, however, summary judgment was granted in favor of OpenAI.[17] The court found that (1) no reasonable reader could have concluded that ChatGPT was communicating the statements about Walters and (2) Walters was a public figure, so he must have shown actual malice, which he could not do.[18] The court in Walters stuck with the traditional framework of defamation law and applied it to this AI case. The Walters case highlights that even when false statements are produced by generative AI, the route towards an equitable outcome is not easy.
These AI defamation cases are not limited to the United States. One such case was initiated in Australia when Mayor Brian Hood alleged that ChatGPT falsely claimed he had been imprisoned for bribery.[19] The lawsuit was dropped after ChatGPT fixed the incorrect allegations.[20] With the continued rise in generative AI use, more scenarios like the ones involving Starbuck and Hood could occur.
In combination with Starbuck v. Meta, Walters will lay the framework for future AI defamation cases. Practitioners need to understand the nuances of these decisions and any changes in the AI and defamation laws to better advocate on behalf of clients harmed by AI-generated false statements.
III. Discussion
How Courts and Practitioners are Approaching AI Defamation Cases
As previously stated, defamation occurs when a false statement, purported to be a fact, is published or communicated to a third party, causing harm to another party. The challenge that arises when AI generates defamatory remarks is determining who should be considered the speaker. The court must determine the mental state of the defendant to measure fault.[21] Obviously, determining a machine’s mental state is impractical, so courts have looked at other ways to determine who is at fault when generative AI produces a false statement.
One way the courts have begun navigating difficult AI defamation cases is by narrowing the reading of Section 230.[22][CL2] Section 230 was designed to shield internet platforms from defamation liability, and courts have begun to consider whether that protection extends to AI companies when the content is generated by the model rather than the user.[23] For example, courts have begun to distinguish between hosting third-party content, which Section 230 protects, and generating the content.[24] While courts shift their handling of defamation cases involving AI, lawyers also need to adapt to determine what is necessary to successfully litigate an AI defamation case.
A group of lawyers from Missouri who specialize in defamation cases believes that regardless of who conveys the false message, the core framework of defamation remains intact.[25] They believe that there are alternative avenues to approach the theory of defamation cases if it involves AI. Some litigators have begun to strategize as to how they would approach a defamation case involving AI. For example, if the attorneys could prove that there are no human beings involved in the output, then an argument could be made that Section 230 would not apply.[26] Benjamin Siders, a technology and IP attorney, believes that courts will be hesitant to grant AI companies a blanket free pass to invoke a Section 230 defense.[27] Therefore, when a case reaches litigation, as seen previously, one of the most important things a plaintiff would have to prove is how much human involvement there was in the generative AI output. Given the difficulty of these cases, individuals should prioritize other ways to curb this threat and decrease the likelihood of being victimized before any harm rises to the level of a legal claim.
Proactive Measures Individuals Can Take to Limit Defamatory AI Outputs
The use of AI presents several challenges for individuals seeking to protect their online image, including verifying that online information about them remains factually correct. For example, Olympic hopeful Taylor Conway had a false statement generated about her, accusing her of sexual assault.[28] Taylor is now at the center of a controversy sparked by generative AI. Preventing the spread of this misinformation would have been much easier than having to correct the false statements after the fact. So how can indivdual mitigate the potential impact of AI defamation?[29]
Whenever an individual posts something online or executes a search, they create a digital footprint that can be used as material for generative AI systems. Educating individuals about the risks of geneartive AI can help people develop a plan to prevent false information from being released and, if it does, mitigate the damage. People must be proactive and regularly monitor their digital footprint in order to quickly rectify any information before false allegations or misinformation go viral.[30] This recommendation is particularly vital for high-profile and wealthy individuals, who also should have a plan in place in case the false information becomes an issue.[31] In addition to individuals being proactive, legislatures need to address the issue and carve out laws that hold AI companies responsible for the distribution of false information.
The Legislature’s Role in Curbing Defamation
Although defamation laws vary by state, the crux of those laws is virtually the same. There must be a false statement made about a third party that causes that person harm. The difficulty with generative AI is determining who made the false statement. As technology evolves, lawmakers must amend existing law or enact new laws that address AI and the party responsible for AI-related defamation.
Since the enactment of Section 230, scholars have argued that many courts have misconstrued or overstated Congress’s intent.[32] Although there has been no reliable pattern of court decisions, the Supreme Court is unlikely to guide lower courts in applying Section 230 to cases concerning generative AI.[33] [CL5] Therefore, it is up to Congress to make the necessary changes. Without change, courts will continue following the ambiguous framework of Section 230 and misconstrue it to fit the evolving technology of AI.[34]
Although the United States courts are dealing with outdated laws, other countries have taken measures to help evolve with technology. The United Kingdom’s Defamation Act and the European Union’s AI Act leave room for plaintiffs to assert that automated systems cause reputational harm.[35] As global regulators tighten the reins on the potential harm stemming from generative AI, the United States must do the same.[36]
With so much confusion, Congress needs to either clarify Section 230 or abolish it and create a new governing statute. To do this, Congress should consider the expansive use of generative AI to reform Section 230 and create a framework that recognizes the evolution in technology.[37] By learning from the United Kingdom, European Union, and other countries that are addressing the same issue, the United States can develop laws that balance both the need of innvoatioon with the responsibility of generative AI developers. Everyone should be held responsible for checking sources and learning to distinguish between false and accurate information. With this in mind, Congress should be able to clarify the governing statute, which will create a solid framework for courts to follow and relieve litigators and courts of confusion[CL6] [AB7] [CL8] .
IV. Conclusion
As technology evolves, the United States needs updated statutes that can both govern emerging technologies and address the growing number of AI-related defamation claims. Currently, courts are attempting to apply traditional defamation laws to fit within the generative AI landscape, which the existing doctrine was not designed to address. This approach has confused lower courts, resulting in the inability to provide concrete or reliable decisions. Consequently, individuals and litigators are left uncertain about how defamation law applies to AI-generated misinformation.
Until Congress takes action, individuals and companies that have become victims of AI-generated misinformation will remain trapped with little recourse. In an era where machines can speak to millions in seconds, the law must be equally swift and deliberate in evolving with the times. This blog suggests that Congress needs to reform the existing laws to help better deal with the advances in technology.
[1] Nat’l Educ. Ass’n, Report of the NEA Task Force on Artificial Intelligence in Education 4 (2024), https://www.nea.org/sites/default/files/2024-10/nea-ai-task-force-report-2024.pdf [https://perma.cc/URC3-EKH8].
[2] Id.
[3] Id.
[4] Id.
[5] Kent F. Hubert et al., The Current State of Artificial Intelligence Generative Language Models Is More Creative than Humans on Divergent Thinking Tasks, 14 Sci. Rep. 3440 (2024), https://www.nature.com/articles/s41598-024-53303-w [https://perma.cc/A5WN-FU7P].
[6] When AI Gets it Wrong: Addressing AI Hallucination and Bias, MIT Sloan EdTech, https://mitsloanedtech.mit.edu/ai/basics/addressing-ai-hallucinations-and-bias (last visited Feb. 22, 2026), [https://perma.cc/LW23-BMTU].
[7] Id.
[8] Defamation, Wex, https://www.law.cornell.edu/wex/defamation (last updated June 2023) [https://perma.cc/FM35-GVDF].
[9] Timothy Prestianni, 131 AI Statistics and Trends for 2026, Nat’l Univ.: Blog (Mar. 4, 2025), https://www.nu.edu/blog/ai-statistics-trends/ [https://perma.cc/4BE9-LGWL].
[10] Id.
[11] Krista Lee Baughman & Amber Hulse, When AI Defames: Global Precedents and the Stakes in Starbuck v Meta, Fed. Soc’y: FedSoc Blog (May 8, 2025), https://fedsoc.org/commentary/fedsoc-blog/when-ai-defames-global-precedents-and-the-stakes-in-starbuck-v-meta [https://perma.cc/8ZLF-VHXY].
[12] Id.
[13] Id.
[14] Id.
[15] Leslie Y. Garfield Tenzer, Defamation in the Age of Artificial Intelligence, 80 NYU Ann. Surv. Am. L. 135, 137 (2024).
[16] Lan P. Vu, AI Defamation Duits Test Anti-SLAPP Shield and Speaker Rights, Daily J. (Dec. 12, 2025), https://www.dailyjournal.com/mcle/1780-ai-defamation-suits-test-anti-slapp-shields-and-speaker-rights [https://perma.cc/W4TX-22FC].
[17] Id.
[18] Id.
[19] Baughman & Hulse, supra note 11.
[20] Id.
[21] Vu, supra note 16.
[22] Id.
[23] Id.
[24] Id.
[25] Eric Achenbach, Courts Test New Frontier of Defamation Law as AI Enters the Mix, MO. Laws. Media (Nov. 12, 2025), https://molawyersmedia.com/2025/11/12/ai-defamation-law-court-cases-openai-microsoft-google/ [https://perma.cc/Z6BM-GMFS].
[26] Id.
[27] Id.
[28] Adele Ashton & Gareth Jones, Guarding Your Reputation: Defending Against AI-Driven Defamation, Penningtons Manches Cooper (Oct. 10, 2024), https://www.penningtonslaw.com/insights/guarding-your-reputation-defending-against-ai-driven-defamation/ [https://perma.cc/5NL4-2A58].
[29] Id.
[30] Id.
[31] Id.
[32] Kathryn Binder, Artificial Intelligence & Defamation Law: An Excuse to Do Away with the Infamously Controversial Section 230?, 76 UC L.J. 1539, 1558 (2025).
[33] Id. at 1560.
[34] Id. at 1558.
[35] Mark Funk & Hope Anderson, Courts Navigating AI Defamation Opens Legal Risks for Companies, Bloomberg L.: Insights & Comment. (Dec. 17, 2025), https://news.bloomberglaw.com/legal-exchange-insights-and-commentary/courts-navigating-ai-defamation-opens-legal-risks-for-companies %5Bhttps://perma.cc/QP9U-CZBF%5D.
[36] Baughman & Hulse, supra note 11.
[37] Dan Ray, When Artificial Intelligence Defames You, Who’s Liable?, NOLO (June 25, 2025), https://www.nolo.com/legal-encyclopedia/artificial-intelligence-defamation-and-libel-is-anyone-liable.html [https://perma.cc/4E86-DFB6].
[CL1]This paragraph could use a topic sentence to tie defamation to the rest of the paragraph. I suggested one…
[CL2]These are cited to the wrong article. Please make sure I assign the right article (this and the Id that follows…)
[CL3]Please go through this section first and read my other comments. I am finding this section to need more fleshing out, with expansion and examples needed. The sentences do NOT always flow under the topic sentence and often lack the logical order. There is great content that is being lost.
[AB4]I went back and reworked some of the sentences. Let me know what you think.
[CL5]When I moved it, I may have messed up the citations.
[CL6]“Counterpoint: Would this approach risk chilling innovation? Because foundational models are trained at scale on diverse data, they often exhibit capabilities not explicitly intended by their developers—reasoning, summarization, coding, etc. If AI companies are held liable for misinformation produced by large language models, even when the model is synthesizing publicly available information in direct response to a user’s request, the incentive to develop and maintain such models could diminish. It may be worth clarifying how liability would apply in contexts where the model is performing as designed and the user’s query drives the output.”
[AB7]I understand your point here. Do you think this should be added as my last paragraph? I am struggling to determine where to put this aspect.
[CL8]Maybe in this paragraph or separating it into another. I think if you talk about what other countries are doing, you can then suggest that the US balance both the needs of innovation with the responsibility of developers and make some suggestions. If true (see comment on my AI- report), I really like that the EU is creating liability for people who repost misinformation. Everyone should be responsible for checking sources and learning to distinguish false information. And if it does chill it a little in the name of safety, then it will be like everything else. We balance ethics and morality with advancement everyday.
The University of Cincinnati Intellectual Property and Computer Law Journal 
Leave a comment