The first instalment of the quarterly Media Futures seminar series was hosted earlier this year at Stellenbosch University by the Department of Journalism (in partnership with the Faculty of Arts and Social Sciences).
The first seminar, titled “The future of intelligence: ChatGPT and its implications”, was chaired by Herman Wasserman (professor of journalism and chair of the Department of Journalism at Stellenbosch University).
Two reports about this seminar, written by BA Honours journalism students, were selected in partnership with Stellenbosch University’s Department of Journalism for publication on LitNet.
Herewith Hannah Theron’s report:
The implications of ChatGPT in academia and journalism were discussed at the first Media Futures seminar held at Stellenbosch University (SU) on Monday, 27 February. The seminar, hosted by SU’s Department of Journalism, included five panellists who provided an overview of what ChatGPT is and the challenges it creates for journalism and teaching.
Media Futures is the start of a new quarterly seminar series intended to invite public engagement, according to Herman Wasserman, the chairperson of the Department of Journalism at Stellenbosch University. “With this seminar series, we want to spark conversations about the media and African societies of the future,” said Wasserman. “The goal of these conversations is also to help us reflect critically on the teaching and scholarship of journalism and media studies in this rapidly changing ecosystem,” added Wasserman.
Five panellists spoke about the implications of ChatGPT for academia and journalism. Photo: Hannah Theron
What is ChatGPT?
ChatGPT is a conversational AI program that was created by a company called OpenAI, according to Bruce Watson from the Centre for Al Research (CAIR). ChatGPT is currently one of the fastest growing apps of its kind, according to Watson.
“[ChatGPT] is certainly one of the most heavily used websites in history,” said Watson. Popular uses for ChatGPT include having it compose essays, requesting political opinions and asking it technical questions, according to Watson.
ChatGPT is able to answer various questions by getting information from a vast number of sources and reshuffling the text, stated Watson. “Unlike many search engines where you just make one query and then it comes back with some answers, if you then actually engage ChatGPT in a bit of conversation, … it will tailor the follow-on questions according to how the conversation has proceeded so far – and, of course, also according to the personal history it’s got with you,” said Watson.
ChatGPT makes use of an artificial neural network that is a digital mimicry of a human neural network, according to Watson. However, it is not sentient and has no hope of developing sentience, stated Watson. “It doesn’t have any intrinsic understanding of the underlying linguistics of English,” said Watson.
The implications of ChatGPT for learning and teaching
ChatGPT is here to stay, and it will only get more powerful, according to Antoinette van der Merwe, senior director of Learning and Teaching Enhancement at SU. The desire to want to either ban or control ChatGPT is the most common initial response to the program, according to Van der Merwe.
Instead of banning ChatGPT, Van der Merwe argues for rethinking and refocusing teaching and learning assessment. “We believe it really … brings very exciting opportunities for us to reimagine our approach to teaching and learning assessment at SU, so we can equip our graduates … with the necessary tools to work responsibly with AI,” she added.
The Learning and Teaching Enhancement Centre views ChatGPT as a partner, stated Van der Merwe. “We should encourage students to use ChatGPT for pure learning and feedback … encourage creating prompt engineering.”
The first Media Futures seminar was hosted by the Department of Journalism at Stellenbosch University on Monday, 27 February. Photo: Hannah Theron
The influence of ChatGPT on journalism
“While these technologies have made it possible for more people to access, participate in and contribute to and curate information, it has also radically changed the way in which journalism is practised and consumed,” said Wim de Villiers, the vice-chancellor and rector of SU, at the start of the seminar.
ChatGPT could assist journalists with acquiring news, or shift data for investigative reporting, stated Fanie van Rooyen, the editor of Quest magazine.
“I think, in simple terms, [ChatGPT] might make some journalistic tasks much easier, and might make discerning the truth more difficult – probably way more difficult – which could make journalism more important than it has been for the past couple of years. [It] could also hold journalism accountable,” said Van Rooyen.
Regulating AI in the public interest
Data quality and systemic discrimination have been highlighted by AI experts as the two main problems with ChatGPT, according to Zara Schroeder, researcher and communications coordinator at Research ICT Africa. “Since its release in November 2022, ChatGPT has already been flagged as being a misogynistic and racist AI system,” said Schroeder.
The AI system is inconsistent and cannot be applied in high stakes situations, particularly ones that deal with marginalised identities, stated Schroeder. “It has a serious issue with robustness. This means the AI system is sensitive to minor typos and misspelt words, which in turn affects the response that is generated by the system,” said Schroeder.
“[ChatGPT is] a very good example of how digital media technology has education for the wider society and cuts across different disciplines,” said Wim de Villiers, the vice-chancellor and rector of Stellenbosch University. Photo: Talia Kincaid
ChatGPT: good or bad?
There are a variety of outlooks regarding ChatGPT, as highlighted by the seminar panellists. “When certain technologies become commonplace, there will inevitably be winners and losers,” said Van Rooyen.
“Futurists are arguing that ChatGPT can be used to provide cheap legal, educational and medical advice for the poor,” said Dr Scott Timcke, senior research associate and communications director at Research ICT Africa.
On the other hand, personality tests administered to ChatGPT by researchers revealed that there was a constant left-leaning personality, which shows that human prejudice can affect AI, stated Van Rooyen. “It incentivises laziness and a ‘copy-paste’ culture. Plagiarism is far easier now,” said Van Rooyen, adding that the “cost of creating fake news is zero”.
However, at the end of the day ChatGPT is a tool. In itself, it is neither good nor bad, stated Van Rooyen. It depends on how humans use it.