Every few weeks, it seems that there is a new headline about a lawyer that has problems with filing filing, in the words of a judge, “Bogus A-generated research”. Details vary, but its line is the same: a lawyer refers to a large language model (LLM) such as Chat to help them in legal research (or worse, written), LLM describes cases that are not present, and the lawyer is not wise unless the judge or the opposing lawyer is not wise. In some cases, including a aviation case since 2023, lawyers have to pay fines for filing filing with AI-infalit deception. So why not stop?
The answer is mostly below the lack of time, and the way AI enters almost every profession. There are now AI integration in legal research databases such as Lexanxes and Westla. For lawyers who awaken the burden of large cases, AI looks like an incredibly effective assistant. Most lawyers do not necessarily use Chat GPT to write their filing, but they are using it and other LLM for research. Yet many of these lawyers, like people, don’t think what the LLM are or how they work. A lawyer who was approved in 2023 said he believes Chat GPT is a “super search engine”. Involed for filing filing with fake references to reveal that it is like a random refrigerator generator-who can either give you the right information or make sure nonsense.
Andrew Pearman, Dean of Safulk University Law School, says many lawyers are using AI tools without any incident, and those who are trapped in fake references are about to go out. Pearlman said, “I think what we are seeing now – though these fraud issues are real, and lawyers have to take it very seriously and be careful about it – this does not mean that these tools do not have too much potential benefits and use to provide legal services.” Legal database and research systems such as Westlaw AI services are incorporating.
In fact, 63 % of lawyers surveyed by Thomson Reuters in 2024 said they had used AI in the past, and 12 % said they use it regularly. The respondents said they use AI to write a case of law and to research “case law, laws, form or sample language for sample language”. The lawyers surveyed by Thomson Reuters have seen it as a time -saving device, and half of the surveyors said that “looking for the possibility of AI’s implementation at work” is their top priority. One respondent said, “The role of a good lawyer is not as a producer of documents as a ‘reliable adviser’.
But as many of the recent examples have been shown, documents produced by AI are not always correct, and in some cases are not real.
In a recent high -profile case, journalist Tim Burke’s lawyers, who were arrested in 2024 for publishing unorganized Fox News footage, presented a motion to dismiss the case on the basis of the first amendment. According to Florida’s Middle District Judge Catherine Kambal Missil, this filing “understood important misrepresentation and misunderstandings and misunderstandings,” Missile found nine deception in the document Tampa Bay Times.
Missile eventually allows Burke’s lawyers, Mark Resich and Michael Maddox to offer a new movement. In a separate filing, explaining the mistakes, Rasch writes that he “takes the sole and special responsibility of these mistakes.” Rasch said he used the “deep research” feature on Chat GPT Pro, which Stuffy Earlier has been tested with mixed results as well as Vestlav’s AI feature.
Rasch is not alone. Anthropic lawyers have recently acknowledged the use of the company’s Claude AI to write a statement presented by music publishers as part of a copyright violation case brought against humanity. This filing also includes a reference with “wrong title and wrong authors”. Last December, Jeff Hanok, a false information expert, admitted that he used Chat GPT to help manage the references in the statement, which he presented in support of the Minnesota law, which managed the use of Deep Fake. Hancock’s filing was “two reference errors, which are popularly called ‘deception,’, and the authors were incorrectly listed for another reference.
These documents are, in fact, of importance – at least in the eyes of the judges. In a recent case, a California judge initially chaired the case against the State Farm through a briefing argument, just to find out that the law of the case was fully made. Judge Michael Wilner wrote, “I read them briefly, he was pleased (or at least interesting) by the authorities, which he cited, and considered the decisions to find out more – just to find out that he was not present.”
Pearlman said lawyers use Generative AI in their work, including finding information in large destinations of discovered documents, reviewing a briefing or filing, and making minds over potential arguments or potential anti -ideology. Pearlman said, “I think there are almost every work, there are ways in which Generative AI can be useful – there is no substitute for lawyers ‘decision, not an alternative to the skills that the lawyers bring to the table, but also enable lawyers’ work and enable them to improve their work, faster and cheaper.”
Pearlman said, but like anyone who uses AI tools, lawyers who rely on legal research and writing on those who need to be careful to test their work. Part of this problem is that lawyers often find themselves short on time – there was a problem before the LLMS came into the picture. “Before the emergence of Generative AI, lawyers used to file documents with references that were not addressed, which they claimed to address,” said Pearlman. “It was just a different type of problem. Sometimes when the lawyers are delivered quickly, they enter the references, they do not check them properly. They do not really see if the case has been overturned or terminated.” (He said, matters are at least common.)
Another, more captaincy problem is that lawyers – like others who use LLM to help research and write – also rely on the birth of AI. Pearlman said, “I think many people are in a sense of relief with the output, because it is so well prepared at first glance.”
Alexander Coloden, a representative of the electoral lawyer and Republican State in Arizona, said he presents the Chattagpat as a junior level partner. They have also used Chat GPT to help write legislation. In 2024, it also added AI’s text in a section of a bill on Deep Fax, LLM provided Deep Fax’s “baseline definition” and then “in human rights protection, such things were included in humor, humor, criticism, criticism, artistic expression, such things,” Coloden said. Benefactor At the time, Coloden said he had discussed the use of Chat GPT with the Central Democratic Caspanser of the Bill, but otherwise he wanted to become an Easter egg in the bill. The bill was passed in the law.
Coloden – who was approved by the Arizona State Bar in 2020 in cases challenging the 2020 elections – also used Chat GPT to write the first draft of the amendment, and reported. Stuffy He also uses it for legal research. He HE to avoid the problem of deception, he said, just checking the references to ensure that they are real.
“You usually do not send a junior associate work products without checking references,” said Coloden. “These are not just machines that a junior associate can read the matter wrong, it does not really stand for a suggestion made in any way. You still have to check it out, but you have to do this with any partner, unless they have experience.”
Coloden said he uses both Chat GPT supporters “deep research” tools and lexins AI tools. Like Westlav, Lexanxes is a legal research tool used by lawyers. In his experience, in his experience, it has a higher fraud rate than Chat GPT, which he says, “has decreased significantly over the past one year.”
The use of AI in lawyers has become so common that in 2024, the American Bar Association issued its first guidance on the use of lawyers from LLM and other AI tools.
The opinion states that lawyers who use AI tools “have the duty of qualification, including maintaining the relevant technical qualifications, which requires an understanding of the developing nature of the Generative A”. Guidance advises lawyers to “get a general understanding of the benefits and risks of GAI tools” that they use – or, in other words, do not think that LLM is a “super search engine”. It states that lawyers should also weigh the dangers of confidentiality of their affairs in LLM and consider whether their clients should tell their clients about the use of LLMS and other AI tools.
Pearlman is happy with the use of lawyers’ AI. “I think the Generative AI legal profession is about to become the most effective technology ever, and that lawyers will be expected to use these tools in the future,” he said. “I think at one time, we will stop worrying about the ability of lawyers who use these tools and start worrying about the lawyers who do not do so.”
The other, including a judge who approved the lawyers to submit a filing filled filled filled files, are more doubtful. Wilner wrote, “Despite the recent developments,” no reasonable lawyer should research and write this technology-especially without any attempt to confirm the accuracy of this content. “
