ChatGPT & Plagiarism: The End of Originality as We Know It?
by Thorsten Beck
The text-based dialog system ChatGPT is the talk of the town. Since the end of 2022, the chatbot has spread rapidly and seems to revolutionize human-machine interaction. However, the ability of the system, trained by self-supervised-learning, to simulate natural language and link meaningful content is seen by more than a few as a provocation and raises a number of questions about values, about originality and plagiarism.
Pupils or students can now, for example, create a sophisticated text at the touch of a button and thus save time on homework. Plagiarism in the conventional sense would no longer be detectable. The general question is to what extent the chatbot will shake the concept of plagiarism or harm good scientific practice? Are the texts generated by A.I. to be considered plagiarism in the strict sense at all?
The conventional definition of the University of Oxford states: „Plagiarism is presenting someone else’s work or ideas as your own, with or without their consent, by incorporating it into your work without full acknowledgement. All published and unpublished material, whether in manuscript, printed or electronic form, is covered under this definition. Plagiarism may be intentional or reckless, or unintentional. Under the regulations for examinations, intentional or reckless plagiarism is a disciplinary offence.“ (https://www.ox.ac.uk/students/academic/guidance/skills/plagiarism)
This definition would declare the use of texts from the chatbot as a "disciplinary offense", if the chatbot can be attributed the status of a subject - but how could such a text transfer by a chatbot be detected? It is obvious that even with an appropriate plagiarism definition, neither the misuse of such a dialog system nor plagiarism can be prevented. Does a chatbot steal intellectual property, or is it rather its purpose to pass on texts? Does our concept of originality fit the new age of artificial intelligence, or does it need to be rethought?
Of course, the curse and blessing of any new technology depends heavily on its use and benefits. Artificial intelligence can detect delays in rail traffic and suggest alternative travel routes to passengers, or can evaluate thermal imaging cameras at airports to prevent the spread of a pandemic. In the case of ChatGPT, a variety of security concerns were voiced. Will the system be instrumentalized for cybercrime, will fishing emails become more effective, will the chatbot - if you bypass its security filters - even pass on plans for building bombs, or explain how to hotwire a car? (https://www.spiegel.de/netzwelt/web/chatgpt-wie-sich-die-kuenstliche-intelligenz-mit-worten-hacken-laesst-a-2a3dd1b4-7405-40e0-8ba0-172915f38e57)
Educators are also expressing concern, such as Adrian Wallbank, writing for Times Higher Education:
„In essence, the embedding of AI (notably ChatGPT and its inevitable successors) in our teaching and assessment will erode student autonomy over their learning experiences, disenfranchise independent thinking and ultimately nullify freedom of expression.“ (https://www.timeshighereducation.com/campus/chatgpt-and-ai-writers-threat-student-agency-and-free-will)
The chatbot, it is argued, is now depriving students of a fruitful learning experience and an important step in the unfolding of a distinct personality - finding one's own voice. Of course, the use of the chatbot should be discussed critically, but doesn't A.I. basically do what people do when they learn: we train with texts and study topics and subjects, follow the literature, and that's how our thinking, our interests, our language are formed. Those who formulate their own ideas and write texts create something new, but foreign thoughts or words enter our thinking as well. The problem that arises with artificially generated texts is that one does not know whether and to what extent foreign words or ideas are recycled. Are phrases taken over verbatim, or ideas copied? The interesting question that accompanies such considerations is: what particular human ability is associated with originality? When is something genuine, independent, and new? Or is not much of what people formulate in terms of creativity and artistic expression also a remix of what has already been created? Kirby Ferguson has shown this impressively in his "Everything is a Remix" video series for the field of "cultural appropriation". (https://www.everythingisaremix.info/) Is the highly valued human capacity for creativity and originality perhaps more of an idealized notion that has no adequate counterpart in the real world?
Perhaps ChatGPT itself should be asked about its influence on the originality or plagiarism debate:
Human: „Write article about how ChatGPT affects the concepts of originality and plagiarism.“
ChatGPT (response): „(…) The use of ChatGPT has had a major impact on the concepts of originality and plagiarism. On the one hand, it has made it easier for people to access information and create content without plagiarizing. With ChatGPT, users can easily create text based on their own conversations with the AI, making it difficult for someone to plagiarize their work. On the other hand, some worry that ChatGPT could be used as a tool for plagiarism. If users were to copy and paste text generated by the AI into their own work, it would be difficult to tell whether the text was originally created by the AI or by the user. Ultimately, ChatGPT has opened up new possibilities for content creation, while also raising questions about the concepts of originality and plagiarism. It is up to users to ensure that they are creating original content, and to avoid plagiarism, by using ChatGPT responsibly.“ (Source: https://chat.openai.com/chat)
Quite obviously, even the A.I. does not seem to be able to distinguish clearly between legitimate text production and its inappropriate use. Basically, texts generated by ChatGPT are similar to texts written by humans. In order to use them, they would have to be citable. But since there is no comparable source (the same queries always generate new variations of responses), the origin of the texts cannot be traced and they cannot be cited in the conventional sense. The door is wide open for misuse. In any case, Pandora's box has been opened and artificial intelligence is on the rise. It will continue to change our lives, our habits and our thinking. Perhaps we will soon find our real counterpart in A.I., as in Spike Jonze's film "Her“ (2013), in which the protagonist Theodore Twombly falls in love with the pleasant voice of an artificial intelligence that seems to know and understand him completely.
Even if one can critically question the concept of creativity and human originality - as long as there are no forensic tools that can detect whether a text was written by a human or a machine, we will have to live with the doubt about the originality of texts. Perhaps it will also be the case that we humans will increasingly become editors of artificial intelligence? After all, many everyday demands and recurring tasks require more routine than originality. The ChatGPT bot writes poems, advertising messages and cooking recipes. It learns formal text structures and can put existing knowledge into new contexts. For such tasks, originality is not necessarily needed. But the chatbot also knows how important originality is for human design and expression:
Human: „How much originality do we need in our everyday lifes?“
ChatGPT (answer): „Originality is important in our everyday lives because it allows us to express ourselves and be creative. We need a certain amount of originality to help us stand out, think outside the box, and be innovative. It can also lead to more meaningful and enjoyable experiences. Originality can also help us to find solutions to problems and create new ideas. So, while a certain amount of originality is necessary for our everyday lives, the amount that is needed depends on the individual and their goals.“ (Source: https://chat.openai.com/chat)
The question of the extent to which artificial intelligence will shape and change knowledge production and the human learning process requires reflection, critical thinking and a lot of discussion. Whether, as Adrian Wallbank believes, the return to a dialogic principle in teaching offers an answer to the challenge posed by advancing artificial intelligence, only time will tell.