Hallucination (artificial intelligence)
In the field of artificial intelligence (AI), a hallucination or artificial hallucination (also called bullshitting, confabulation, delusion, or mirage ) is a response generated by AI that contains false or misleading information presented…
≈ 9 min read
Names in other languages
No names in other languages are held for this entry yet.
Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.
In the field of artificial intelligence (AI), a hallucination or artificial hallucination (also called bullshitting, confabulation, delusion, or mirage ) is a response generated by AI that contains false or misleading information presented as fact. These terms draw a loose analogy with human psychology, where a hallucination typically involves false percepts.
Chatbots powered by large language models (LLMs), like ChatGPT, may embed plausible-sounding random falsehoods within their generated content, such as fabricated citations. Detecting and mitigating errors and hallucinations pose significant challenges for practical deployment and reliability of LLMs in high-stakes scenarios, such as chip design, supply chain logistics, and medical diagnostics.
Some software engineers and statisticians have criticized the specific term "AI hallucination" for unreasonably anthropomorphizing computers. Symbolic artificial intelligence models generally do not produce hallucinations, unlike large language models.
Causes
In natural language generation, there are several reasons why LLMs hallucinate. The pre-training of generative pretrained transformers (GPT) involves predicting the next word, which incentivizes them to "give a guess" about what the next word is, even when they lack information. Hallucinations can also stem from incomplete, inaccurate or unrepresentative datasets.
Some researchers take an anthropomorphic perspective and posit that hallucinations arise from a tension between novelty and usefulness. For instance, Amabile and Pratt define human creativity as the production of novel and useful ideas.
By extension, a focus on novelty in machine creativity can lead to the production of original but inaccurate responses—that is, falsehoods—whereas a focus on usefulness may result in memorized content lacking originality. In 20251, according to interpretability research by Anthropic on Claude, the LLM appears to have internal circuits that cause it to decline to answer questions unless it knows the answer.
By default, the circuits are active and the LLM doesn't answer. When the LLM has sufficient information, these circuits are inhibited and the LLM answers the question.
The researchers said that hallucinations were found to occur when this inhibition happens incorrectly, such as when Claude recognizes a name but lacks sufficient information about that person, causing it to generate plausible but untrue responses.
United States
The adoption of generative AI by law firms for legal research during the 2020s resulted in many submissions to courts containing hallucinations, with U.S. judges nationwide sanctioning or reprimanding lawyers. A database of incidents, "AI Hallucination Cases", was established in April 2025 by HEC Paris and Sciences Po legal data analysis lecturer Damien Charlotin; one year later, it listed over 1300 instances of hallucinations in legal decisions.
In May 2023, it was discovered that Stephen Schwartz had submitted six fake case precedents generated by ChatGPT in his brief to the Southern District of New York on Mata v. Avianca, Inc., a personal injury case against the airline Avianca. Schwartz said that he had never previously used ChatGPT, that he did not recognize the possibility that ChatGPT's output could have been fabricated, and that ChatGPT continued to assert the authenticity of the precedents after their nonexistence was discovered.
In response, Brantley Starr of the Northern District of Texas banned the submission of AI-generated case filings that have not been reviewed by a human, noting that: On 23 June 2023, judge P. Kevin Castel dismissed the Mata case and issued a $5,000 fine to Schwartz and another lawyer—who had both continued to stand by the fictitious precedents despite Schwartz's previous claims—for bad faith conduct. Castel characterized numerous errors and inconsistencies in the opinion summaries, describing one of the cited opinions as "gibberish" and "[bordering] on nonsensical".
In May 2025, high-profile pro se plaintiff Ashley Gjøvik, who has a Juris Doctor from Santa Clara University School of Law, was admonished in U.S. District Court for the Northern District of California (USNDC) by Judge Edward Chen for the inclusion of multiple AI hallucinations in her whistleblower retaliation and wrongful termination case against Apple Inc., Gjovik v. Apple Inc., warning that future instances would result in Gjøvik being held in contempt of court and her pro se privileges revoked. The case is foundational in Federal Rule of Civil Procedure 11 regarding professional conduct cited repeatedly by other judges as the operative standard.
It has also been discussed by the American Bar Association in updating the model professional rules and cited in its opinion regarding the use of generative AI. In September 2025, USNDC Magistrate Judge Peter Kang sanctioned Webb Law attorney Katherine Cervantes for her submission of July 21 that included an AI hallucination which she stated had been obtained from Westlaw AI, the Thomson Reuters' artificial intelligence tool.
On April 28, 2026, Judge Kang sanctioned and fined Webb Law Group managing partner Lenden Webb for insufficient supervision of Cervantes—a junior lawyer incorporating AI for the first time—therefore allowing the legal submission that included an AI hallucinated citation of July 21, 2025. On 18 March 2026, Clayton County, Georgia, Assistant District Attorney Deborah Leslie appeared before the Supreme Court of Georgia to deliver oral arguments regarding an order to deny an appeal for a murder conviction, which a Clayton County Superior Court judge used in denying the defendant's motion.
During the session, Chief Justice Nels S. D. Peterson noted at least five fictitious citations within the filing, as well as five others that did not support the proposition for which they were cited. On 30 March 2026, Leslie admitted in an affidavit to not verifying AI-generated "expanded legal research" used to prepare the order, which District Attorney Tasha Mosley corroborated in a written apology to Peterson.
On 5 May, the Supreme Court vacated the lower court's ruling and ordered for a new order with legitimate citations to be written. Leslie was also suspended from practicing before the Supreme Court for six months and was ordered to take additional education for ethics and proper use of artificial intelligence.
In April 2026, Sullivan & Cromwell submitted an apology to the U.S. Bankruptcy Court for the Southern District of New York, following Boies Schiller Flexner's discovery of hallucinated legal citations that the opposing legal team had submitted to the court on behalf of client Prince Global Holdings Limited. On 28 April 2026, U.S. Magistrate Judge Robert T. Numbers formally reprimanded former assistant U.S. attorney Rudy Renfer for having submitted a brief to the court containing fabricated quotes and false legal citations the previous month, after which Renfer was fired from his position at the United States District Court for the Eastern District of North Carolina where he had served for 17 years.
Israel
As of August 2026, Israeli courts and tribunals had identified at least 55 cases in which parties submitted documents containing fictitious citations or otherwise AI-hallucinated arguments. On February 2025, the Israeli High Court of Justice warned litigants that further incidents of this kind would attract sanctions such as personal cost orders or disciplinary proceedings before the Bar.
The emerging Israeli case law on AI hallucinations also included instances in which police prosecutors and municipal authorities relied on fictitious statutory provisions.
England and Wales
The High Court of Justice criticised the use of generative AI-based pleadings with hallucinated authorities in the conjoined 2025 case of Ayinde v Haringey and Al-Haroun v Qatar. In addition, the UK's First-tier Tribunal and Upper Tribunal has dealt with a number of cases where litigants have generated hallucinated legal citations.
Legal professional bodies have put out guidance on the use of AI including the Solicitors Regulation Authority, the Law Society of England and Wales, and the Bar Council.
Mitigation methods
As of 2025, there is still ongoing research to try to mitigate the occurrence of hallucinations. Research in 2022 has shown that language models not only hallucinate but also amplify hallucinations, even for those which were designed to alleviate this issue.
Researchers from OpenAI wrote that hallucinations occur because the training and evaluation of LLMs reward guessing over acknowledging uncertainty, and proposed modifying the scoring of benchmarks. Researchers have proposed a variety of mitigation measures, including getting different chatbots to debate one another until they reach consensus on an answer and neuro-symbolic architectures with formal knowledge bases and inference engines.
Ji et al. divide common mitigation methods into two categories: data-related methods and modeling and inference methods. Data-related methods include building a faithful dataset, cleaning data automatically, and information augmentation by augmenting the inputs with external information.
Model and inference methods include changes in the architecture (either modifying the encoder, attention, or the decoder in various ways); changes in the training process, such as using reinforcement learning; and post-processing methods that can correct hallucinations in the output. Nvidia Guardrails, launched in 2023, can be configured to hard-code certain responses via script instead of leaving them to the LLM.
Furthermore, tools such as the Trustworthy Language Model have been created to attempt to aid in the detection of hallucination. Evaluating multiple possible replies before answering a query by assigning confidence scores to each could mitigate the problem. However, this approach would multiply computational costs.
Active learning would further increase these costs. In high-stakes domains such as chip design, supply chain logistics, and medical diagnostics, the added costs would be operationally necessary in order for LLMs to be used. In chatbots, however, customers tend to prefer rapid, overconfident answers over cautious, uncertainty-aware ones.
Retrieval-augmented generation, used by many modern chatbots, reduces hallucinations by allowing LLMs to access external sources before providing information.
See also
AI alignment AI effect AI safety AI slop Artifact Artificial stupidity Chatbot psychosis Memetic algorithm Turing test Uncanny valley
Text derived from https://en.wikipedia.org/w/index.php?oldid=1378683911 under CC BY-SA 4.0; modifications have been made.
References
- ↩Wikipedia “Hallucination (artificial intelligence)” date of establishment (×1) · confidence 0.95 revision URL