The way in which the media portrays ChatGPT's coverage
The Tow Center conducted an analysis of how news organizations have been reporting on generative AI, specifically focusing on the coverage of ChatGPT over the past six months. With the rapid advancements in AI tools, journalists face the challenge of providing accurate and contextualized reporting to audiences who may be encountering this technology for the first time.
However, some of the coverage has been sensationalist, as highlighted by linguist and social critic Noam Chomsky in a New York Times op-ed. Numerous hyperbolic headlines have emerged, such as "Bing's A.I. Chat: 'I Want to Be Alive.'" or "'Godfather of AI' says AI could kill humans and there might be no way to stop it." These headlines create a sense of alarm and fear.
To gain a better understanding of how ChatGPT is being covered by newsrooms, the Tow Center interviewed academics and journalists to examine how the media has framed the coverage of generative AI chatbots. They also collected data on the volume of online news coverage using the Media Cloud database and analyzed TV news coverage through the Internet TV News Archive.
The coverage of new technologies often follows a hype cycle, according to Felix M. Simon, a researcher at the Oxford Internet Institute. Initially, there are high expectations and promises associated with the technology. This is followed by extreme narratives, portraying either the impending apocalypse or utopian possibilities brought about by the technology. Eventually, the coverage transitions into a more nuanced phase, discussing real-world impacts and moving away from catastrophic or idealistic narratives.
However, the coverage of generative AI chatbots like ChatGPT shows no signs of cooling off. ChatGPT gained significant attention when it was publicly launched by OpenAI in November 2022, with over a million users in just a few days. The media's focus on ChatGPT intensified in January and February 2023 when BuzzFeed announced its use of ChatGPT for content creation, Microsoft integrated a ChatGPT-powered chatbot into Bing, and Google introduced its own challenger, Bard.
One recurring issue with media coverage of this technology, according to Subramaniam Vincent from the Markkula Center for Applied Ethics, is that it tends to rely heavily on the claims made by companies about what the technology can do. This structural problem extends beyond ChatGPT, and company CEOs often take to social media to shape the narrative surrounding AI.
Television news coverage of chatbots from OpenAI, Google, and Microsoft became more frequent in early 2023, as indicated by data from the Internet TV News Archive. Business news channels, particularly CNBC, covered these companies' activities related to generative AI chatbots extensively. Cable news channels had less coverage compared to business news, with CNN and Fox News featuring ChatGPT more prominently than MSNBC. Topics explored by CNN and Fox News included the impact of generative AI on education, the workplace, jobs, and concerns about political bias. Fox News also frequently mentioned Elon Musk, who expressed concerns about ChatGPT becoming "woke" and called for a hiatus in AI tool development.
Accurate news coverage is crucial as many voters are unfamiliar with AI programs like ChatGPT, as revealed in an April Fox News poll. Some news coverage, both on TV and online, has provided nuanced information to help audiences navigate the technology, identify potential issues like hallucinations, and fact-check AI-generated statements. Reporters have also delved into topics such as algorithmic bias, ethical considerations, the spread of misinformation, and the need for regulations to prevent misuse. However, some coverage continues to rely on science fiction-like narratives, fueling fear and uncertainty rather than promoting understanding.
Is it a Cycle of Hype?
While ChatGPT appears to be ushering in a new era, there are similarities in the media coverage to the hype surrounding Bitcoin and cryptocurrencies. Data from Media Cloud's news database indicates that within just six months of its launch, ChatGPT is receiving a comparable level of attention to what cryptocurrencies received in 2021, when Bitcoin prices reached their peak, more than a decade after its public release in 2009.
Some observers have expressed dissatisfaction with the media coverage, acknowledging the presence of a hype cycle. Paris Martineau, a tech reporter at The Information, points to structural challenges in journalism such as declining advertising revenue, reduced budgets, and smaller newsrooms, which contribute to the sensationalized coverage driven by the need for web traffic. Martineau believes this trend reflects the current state of the publishing and news system, where newsrooms feel compelled to cover every angle, including sensationalist ones, to capture audience attention.
Comparing ChatGPT's coverage to other new technologies, it is evident that ChatGPT has received more media attention than technologies like virtual reality and deep fakes, although coverage of cryptocurrencies still surpasses it, particularly after the collapse of FTX.
One possible reason for the extensive coverage of ChatGPT is its direct implications for journalism, as suggested by Jenna Burrell, director of research at Data & Society. Another reason could be the potential of ChatGPT and generative AI tools to disrupt various creative industries.
However, Burrell is concerned about the framing of much of the reporting, emphasizing the need for clarity about how these technologies work and combatting exaggerated hype. The anthropomorphic tendencies in attributing human-like capabilities to non-human tools, such as the claim that Bing's chatbot wanted to "be alive," present a distorted, science-fiction narrative that hinders meaningful discussions around ethics, usage, and the future of work.
Nick Diakopoulos, an associate professor at Northwestern University, describes this as the "Hollywood-ification" of the public's understanding of AI, perpetuating entertaining views without critical examination. Ideally, news coverage should take a more critical and nuanced approach rather than reinforcing these preconceived notions.
Striving for Improved Reporting: Enhancing Coverage of ChatGPT
Towards Improved Media Representations: Shifting Towards Responsible Coverage of Generative AI
To achieve better media representations of generative AI going forward, it is crucial to move beyond the current hype phase. According to Burrell from Data & Society, a more grounded approach is needed to address significant issues that often go overlooked. One such issue is the concentration of power and wealth among a select few who have invested in and built these tools, poised to reap significant financial benefits. It is essential to shift the focus away from sensationalized aspects like AI's potential "sentience" and instead examine how these technologies contribute to the consolidation of wealth and power.
In order to cut through the haze of science fiction, there is a need for more informative reporting on the functionality and limitations of generative AI tools like ChatGPT. Diakopoulos emphasizes that these tools, while proficient at predicting the next word in a sentence based on statistical distributions, do not "think" like humans. Reporting should provide a clear understanding of how these technologies work and highlight their shortcomings.
Newsrooms should address important questions regarding who has the authority to train these models and what biases and flaws may be ingrained within them. Furthermore, editors should reconsider whose perspectives on generative AI are deemed newsworthy.
Simon from the Oxford Internet Institute asserts that sensationalized coverage diverts attention from more critical questions. These questions include newsrooms' potential future dependence on tech giants for news production, the governance decisions made by these companies, ethical considerations and biases related to models and training, the environmental impact of these tools, and more. The aim should be to engage a broader public in contemplating these issues, rather than leaving them solely in the domain of engineers and policy experts.
Establishing guidelines, possibly through style guides, can help newsrooms shape their coverage strategies moving forward. Martineau suggests specific parameters, such as avoiding anthropomorphizing chatbots. Such measures can contribute to cooling the fires of the current hype cycle.