The discussion about ChatGPT technology
The program ChatGPT, which appeared at the end of the last year, opened up new opportunities for creating texts, searching for information on the Internet, and at the same time raised new challenges and questions related to the work of lawyers and technologies. In the discussion held on March 22, 2023 at the VU Faculty of Law, organized by lect. dr. Neringa Gaubienė, the most important features and perspectives of this technology were presented.
At the beginning of the event, Associate Professor of Mathematics and Informatics of VU and member of the board of the Lithuanian Association of Artificial Intelligence dr. Linas Petkevičius presented how large language models work, which were created after using many texts and are intended to predict further text. One such model is ChatGPT. The researcher named the advantages of this program for lawyers: faster obtaining of potentially useful information, preparation of summaries, obtaining basic information, finding opposing arguments and preparing the text in the desired forma.
Another participant in the discussion, Ai expert and head of Dip-Tech Dalius Mašalas raised the question – how can a young employee become a good specialist? During his presentation, it was mentioned that the biggest challenge of ChatGPT technology is for a young professional, and for an experienced legal professional, ChatGPT does not create competition. According to D. Maršal, artificial intelligence will become our assistant and thus we will be able to do better, more creatively, faster. ChatGPT is the first widely available tool that shows how much such an assistant can do. With the arrival of disruptive technology, we need to adapt and consider how we will learn and work in the future.
Irina Seniut, an innovation expert from the „Sorainen“ law firm, presented the effectiveness of ChatGPT in the work of a lawyer. The presentation presented the questions asked by ChatGPT and the answers received. When asked how this technology can help lawyers manage information, ChatGPT generated a response that mentioned the efficiency with which it can process large amounts of data, search for information, as well as the spectrum of information, which means that the system can provide a lot of information about many questions However, the system was described as “self-critical” as it immediately stated that the data it generates could be inaccurate and unreliable. It has also been noted that this software is context-aware, so answers are only given based on the data available. Language imperfection has also been singled out as one of the drawbacks of this system.
The emergence of ChatGPT has brought suspicion to the field of education – how can we trust the uniqueness of written works and answers submitted by schoolchildren and students? The lecturer of the Faculty of Economics and Business Administration dr. Julia Savicke. She identified indications that certain answers to open-ended questions on the exam were prepared with the help of ChatGPT. One of the signs is a recognizable writing style: correct paragraph structure (with introduction, teaching, conclusion), frequent repetition of key words, grammatically correct language. At the same time, a new challenge arises for teachers – to come up with more creative ways to test students’ knowledge and abilities.
During the discussion, the question of protection of intellectual property and private data was raised – it remains unclear whether such data will be used in ChatGPT technology.
