Global Gleanings # 16: More disruption: chatbots in the news

This quarterly column of news, views and snippets from the international literature of books, libraries, and information, appears in LIASA-in-Touch, the newsletter of the Library and Information Association of South Africa.

Generative artificial intelligence

Image: ChatGPT Hub, https://www.chatgpthub.com.br/quem-e-dono-do-chatgpt/

What is ChatGPT? Image: ChatGPTHub, https://www.chatgpthub.com.br/quem-e-dono-do-chatgpt/

What is ChatGPT?At time of writing the big tech news is about ChatGPT-4, the latest version of a generative artificial-intelligence (AI) chatbot launched in November 2022 by a OpenAI, a partly non-profit and partly for-profit artificial intelligence research laboratory based in San Francisco, California. Elon Musk was a co-founder, but has in the meantime pulled out. Microsoft made a large investment in this project, whilst other dominant IT companies such as Google and Amazon are rushing to bring out their own generative AI systems. It has generated much excitement and some apprehension.

What is a generative AI system? Who better to explain it than an AI chatbot, as in this article.  Basically, it is an artificial intelligence algorithm that, having been “trained” using an enormous corpus of text and other content, can generate new outputs. It does this in seconds, producing text, images or music, that can be difficult to distinguish from those produced by us, much slower humans. As they do it repeatedly, these chatbots learn and become more convincing. This has all sorts of implications. Given some parameters and a large enough corpus, they could write blurbs to be pasted on wine bottles, adverts for a new cars or toothpaste, create poems, write student essays, grant proposals, scholarly articles and theses, compose music, and generate images. This will be very useful for many users, and a headache for a host of occupations: copywriters, teachers, professors, and journal editors – and librarians?

ChatGPT and libraries

It was Denise Nicholson’s eBulletin, Scholarly Horizons that alerted me to this topic. Despite our undeserved old-fashioned image, librarians are quick to assess and implement new technologies; a basic Google search turned up various recent articles explaining what ChatGPT is and exploring its potential impact on libraries generally (Lund and Wang 2023) or on specific library services, such as reference service (Chen 2023). An article in College & research libraries news outlines eleven areas of its impact in academic libraries. These include discovery and search, research, teaching, textbooks, information literacy and digital literacy, writing and creation, plagiarism, and copyright (Cox and Tzoc 2023). In this column I focus on the last two of these. They are discussed in academic and scholarly circles, which many libraries serve.

We teach students to acknowledge their sources. If they do not, we penalize them for plagiarism. Some enterprising scholars, having used ChatGPT-4, in composing their draft articles, experimentally “did the right thing” and listed ChatGPT as a co-author. When they submitted these articles to the prestigious journals Nature and Science, they were turned down and the journal publishers quickly updated their editorial policies to forbid this (Staiman 2023). However, Staiman pointed out that AI tools such as Chat-GPT may be very useful for authors for whom English is a second or third language, helping them to formulate their ideas in the form and style of communication expected in high-ranking journals. AI could conceivably help scholars everywhere to cope with the ever-increasing pressure to “publish or perish” (Carrigan 2023).

If the system can write essays, chances are that some students, having relied heavily or entirely on Chat-GPT will not cite it as a co-author of their essays or theses. This is clearly plagiarism. It will become more difficult for instructors to detect it, and this has become a hot topic in higher education (e.g. Halaweh 2023; Khalil and Er 2023). At least a dozen plagiarism detectors have appeared on the market, claiming to be able to detect it. In a test of five of these to determine how successful they were in distinguishing text generated by ChatGPT from text written by a human, some performed better, but none was entirely dependable (Leong 2023).

Plagiarism and infringement of copyright should not be confused. Mostly, plagiarism also entails infringement of copyright. Interestingly, it can be argued that both the creation and use of a chatbot such as ChatGPT inevitably involves infringement of copyright. For its “training” a chatbot needs an enormous corpus of text, audio, and images. This material, is “scraped” from the web by robot programs not unlike those used by Google. All this material is copied, without permission of copyright holders. If you write an article or essay using ChatGPT, you may well be using copyrighted material without permission. Furthermore, the author of a text, a song or an image produced with the help of a chatbot will own the copyright on it, although it has been ruled in the USA and the UK that non-human creators cannot own copyright (Hunter et al. 2023; Presenti 2023). This is as of now. We may be sure that copyright lawyers will have rich pickings sorting it all out. We can also be sure that this disruptive technology will transform art, the music industry and  copyright itself (Merkley 2023).

And libraries? We will no doubt also be affected, as we were by the coming of copying machines, computers, online information retrieval, and the Internet. But we will be busier than ever helping and teaching our clients to make optimal use of chatbots, to avoid and detect plagiarism, and to navigate copyright issues, among the many other challenging tasks which make our profession so satisfying.

“The best things in life are free”

In case you sometimes feel that what we contribute to society is not appreciated, here’s an article that appeared in the Daily Maverick on 21 March to celebrate National Library Week (Dall 2023). Journalist and author Nick Dall describes librarians as “invariably knowledgeable, friendly and helpful.” Are we?  Anyway, humans are more fun than bots. Maybe he should be invited to speak at the next LIASA Conference?

 

References

Carrigan, Mark. 2023. “Could AI Free Academics up or Increase the Pressure to Publish?” University World News. 2023. https://www.universityworldnews.com/post.php?story=20230317075632709.

Chen, Xiaotian. 2023. “ChatGPT and Its Possible Impact on Library Reference Services.” Internet Reference Services Quarterly 0 (0): 1–9. https://doi.org/10.1080/10875301.2023.2181262.

Cox, Christopher, and Elias Tzoc. 2023. “ChatGPT: Implications for Academic Libraries,” March. https://doi.org/10.5860/crln.84.3.99.

Dall, Nick. 2023. “The Best Things in Life Are Free: In Praise of Our Libraries.” Daily Maverick. March 21, 2023. https://www.dailymaverick.co.za/opinionista/2023-03-21-the-best-things-in-life-are-free-in-praise-of-our-libraries/.

Halaweh, Mohanad. 2023. “ChatGPT in Education: Strategies for Responsible Implementation.” Contemporary Educational Technology 15 (2): ep421. https://doi.org/10.30935/cedtech/13036.

Hunter, Chris, Lindsay Toth, Marko Trivun, and Chantelle Hospedales. 2023. “Does Generative AI Need to Infringe Copyright to Create?” JD Supra. March 10, 2023. https://www.jdsupra.com/legalnews/does-generative-ai-need-to-infringe-2221581/.

Khalil, Mohammad, and Erkan Er. 2023. “Will ChatGPT Get You Caught? Rethinking of Plagiarism Detection.” arXiv. https://doi.org/10.48550/arXiv.2302.04335..

Leong, Aaron. 2023. “How to Detect ChatGPT Plagiarism, and Why It’s so Difficult.” Digital Trends. January 20, 2023. https://www.digitaltrends.com/computing/how-to-detect-chatgpt-plagiarism/.

Lund, Brady D., and Ting Wang. 2023. “Chatting about ChatGPT: How May AI and GPT Impact Academia and Libraries?” Library Hi Tech News ahead-of-print (ahead-of-print). https://doi.org/10.1108/LHTN-01-2023-0009.

Merkley, Ryan. 2023. “On AI-Generated Works, Artists, and Intellectual Property.” Lawfare (blog). February 28, 2023. https://www.lawfareblog.com/ai-generated-works-artists-and-intellectual-property.

Presenti, Eran. 2023. “AI and Its Implications for Copyright -Opinion.” The Jerusalem Post | JPost.Com. March 5, 2023. https://www.jpost.com/business-and-innovation/tech-and-start-ups/article-733356.

Staiman, Avi. 2023. “Academic Publishers Are Missing the Point on ChatGPT.” The Scholarly Kitchen (blog). March 31, 2023. https://scholarlykitchen.sspnet.org/2023/03/31/guest-post-academic-publishers-are-missing-the-point-on-chatgpt/.

Unknown's avatar

About Peter Lor

Peter Johan Lor is a Netherlands-born South African librarian and academic. In retirement he continues to pursue scholarly interests as a research fellow in the Department of Information Science at the University of Pretoria, South Africa.
This entry was posted in Copyright, Uncategorized. Bookmark the permalink.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.