Artificial intelligence (AI) is the phrase on everybody’s lips. But why? A guest paper by American researcher Meredith Whittaker on netzpolitik.org provides a plausible explanation: It is a narrative. She claims that the cognition and computer scientist John McCarthy came up with the term in the 1950s in order to attract better funding for his research – and to marginalize unwelcome competition, specifically the information technology expert Norbert Wiener. That is why the more imaginative term »artificial intelligence« was able to assert itself over »cybernetics.«
The narrative of »artificial intelligence,« which can be used as a vessel for a wide range of content, is actually an example of successful public relations for science, Whittaker argues. In her readable and worthwhile article, she highlights the invisible human work behind the enormous quantities of data used to train generative AI platforms. And she also warns of the dangers of the platforms’ »extractive logic.« Society needs to be aware that the data base and the algorithms that process it are based on existing texts, with all their strengths and weaknesses.
In his paper »New game, new rules,« Kim Björn Becker examines these weaknesses in the use of AI in journalism. He was interested to explore the question of how German-language media and press agencies are currently handling AI. Becker successfully shows that »although media are already looking at fundamental questions thrown up by the new technology, newsrooms still have blind spots when it comes to dealing with AI.« It is particularly common for media to ignore the »algorithmic bias« – just two of the media he looked at appeared aware of the problem.
AI companies themselves are above all companies with economic interests, as Meredith Whittaker points out. In relation to newspaper companies, Karl Bücher criticized the subordination of journalistic quality to advertising revenue as the »true seat of the evil« as far back as 1926, as Horst Pöttker shows in his paper on »Karl Bücher’s reform concept.«
According to Bücher, this concept means that anything that »could damage« the advertising section, the elixir of life for the newspaper, is »by default excluded from being covered« in the journalistic section. In order to break this economic dependency on advertising revenue, Pöttker proposes expanding the organizational principle of public service broadcasting to other sectors of the media. Even Karl Bücher had imagined a public advertising monopoly that could separate advertising and journalism in both economic and organizational terms.
Is this feasible? »The approach proposed here would not be as far removed from the current situation as it might first appear,« wrote Bücher. But the narrative of media companies fighting for their economic survival (while bravely maintaining journalistic standards) and having to assert themselves against public service media who are allegedly harmful to competition is widely circulated – by who, we wonder?
Narratives of a very particular kind are the topic of Fox News’ website, foxnews.com. Numerous content analyses have attempted to work out what defines its particular populistic content. In his discourse analysis »Fear and balanced. The world according to foxnews.com homepage,« Fred Vultee also examines the narratives that did not take place on Fox News. Fox News is already outside the mainstream when it comes to negativity in selection and authority in sourcing. For example, there are no cases in which Twitter explodes in favor of a policy aimed at reducing carbon emissions; the vice president is never a figure of policy, only a figure of ridicule; presidential approval is forever plummeting, even when it is rising.
Instead, it is dominated by narratives that confirm the world view of large parts of its audience: Critical race is to be forced onto children in elementary schools, presidential fecklessness is moving China closer to world dominance by the day, cities will collapse under the self-induced burdens of crime and homelessness – the end of the world is nigh. »Whichever side is winning, there is no respite from the ultimate battle of good and evil.« The repetition of this kind of narrative and the way in which it is combined with other pieces are reminiscent of Yulia Belinskaya’s piece in edition 1/2023, in which she described the theory of securitization in relation to the internet in Russia.
Fred Vultee’s discourse analysis is based on a database of foxnews.com homepages from the years 2022–23. It is only when they are put in context that the links between the narratives’ meanings become clear.
Meredith Whittaker’s piece was published after I had already completed my workshop report on ChatGPT in journalism teaching, so I was unable to address the narrative behind it more explicitly. I merely gently mock the predictability of the discourse. But my main focus was on asking myself what journalism teaching looks like in a world of generative AI platforms. Which competencies, which specific knowledge and skills need to be taught?
Together with students, I explored the use of generative language AI in text production during degree studies and reflected on what we had learned. My choice of topics for this exploration was guided by learning objectives, which of course need to be expanded. Examples include checking facts, knowing and applying the journalistic rules of separation, transparency, and reporting and reportage using prompting in the dialog.
We were able to disprove one narrative: Students certainly do not use the tools uncritically, but instead in a well-considered and critical way.
Journalists should recognize narratives. But for various reasons, they often do not. They are particularly vulnerable to a narrative when it is about a good cause. Georg Cremer, long-time Secretary General of the charity Caritas-Verband, has examined the narrative of decline in social policy. In his debate piece »A lack of critical corrective,« he uses various examples to demonstrate how charities’ public relations departments achieve major coups – which does not always help to maintain the debate’s objectivity. His conclusion? »The media frequently fail to provide a critical corrective, especially when press releases from social organizations are in line with preconceived perceptions and the demands that the organizations derive from them appear to be for a good cause.« A lack of statistical skills plays a role here, but so too do priming and framing. Cremer does not use these terms, but describes the situation and its consequences accurately. He would like to see professional training for journalists »encourage openness and curiosity among trainee journalists, so that they develop an interest in exploring whether the development of the situation in society might be different, more complex or more contradictory than they had previously imagined.«
There is no better way to describe the requirements of journalism: to recognize narratives as such and to compare them with the facts as well as possible.
We are inviting submission for the upcoming issues of Journalism Research/Journalistik for Winter 2023 (Issue 3/4, 2023) and Spring 2024 (1/2024) on the special topic of broadcasting in all its forms and manifestations, including public-service broadcasting, private commercial/corporate broadcasting, and last, but not least, non-profit and community broadcasting. Contact email@example.com.
Translation: Sophie Costella
1 Whittaker, Meredith (07. 06 2023): Künstliche Intelligenz: Vermessung bis ins Innerste. Accessed on 12 June 2023 at netzpolitik.org: https://netzpolitik.org/2023/kuenstliche-intelligenz-vermessung-bis-ins-innerste/#netzpolitik-pw This text is based on the speech Meredith Whittaker gave at re:publica 2023.