Does AI jeopardize journalistic credibility?

Prof. Dr. Claudia Paganini, Lena Jakat: Readers can see how easy it is to create texts and images using artificial intelligence. Is journalistic credibility suffering as a result?

Paganini: The new media are not dysfunctional in themselves. They just make visible what was already problematic in communication before. When we talk about AI-generated images, for example, we should have thought about their credibility earlier. An image is always a work of interpretation, but the viewer usually assumes eyewitness quality.

Jakat: In my experience, it can unfortunately happen from time to time that archive and symbol images are used incorrectly in the everyday stress of local newsrooms. AI offers an opportunity here. Firstly, to make it more transparent where there are no original images. And secondly, to visually enrich something. We don't always need the same blue light for police reports. It is important to label AI-generated content in order to convey journalistic credibility.

AI-generated content: Transparency instead of truthfulness as a guiding value

How should AI-generated content be labeled?

Jakat: In future, we should both show that content was created with AI and name the person controlling the AI. Then it will also be clear that there is always human-machine interaction. And that human editors also control the content.

Paganini: I think we should focus much more on transparency as a guiding value than, for example, truthfulness. It should be clear that there is a person in the background who is in charge and tangible. The public still has some catching up to do in terms of media literacy. A lot has to happen, however, so that a government, for example, becomes aware of how important it is to invest in this.

In the US election campaign, an AI-generated flood of fake news is expected to defame political opponents. Can such content permanently damage the public's trust in the media?

Paganini: First of all, I think that this will lead to uncertainty. And that is dangerous because this uncertainty can be exploited to plant conspiracy myths. Of course, it is a problem when technical innovations increase the risk of manipulation. But this already existed without AI-generated content. We need trust, which is why we need to answer the question of how we can give people the sense to recognize what solid sources are. Quality media comes at a price, but people can be sure that they are well informed.

So do we need an "AI free" seal?

Jakat: I'm not so pessimistic about the spread of conspiracy myths. In times of crisis, such as during the pandemic, we have seen that the large, reputable media houses have seen an upsurge. Of course, there is a section that is moving away from democratic institutions. And we are concerned that this part is growing. We need to do more to promote journalism in a positive sense and communicate the contribution we make to a democratic society.

Paganini: I would like to reinforce that, because we also see in various youth studies that young people in particular trust quality media. That's a good signal, because the trust is there.

Skills shortage in journalism: using AI to make work more efficient and attractive

Man-made journalism is being upgraded and we need a strong human corrective. Does this answer the question of whether AI is taking away jobs?

Paganini: The problem may be less serious in journalism. This is because technical innovations generally replace activities for which only a low level of qualification is required.

Jakat: It is illusory to think that AI does not increase efficiency. However, the shortage of skilled workers has long since reached journalism. If we avoid redundant activities, it makes the job more attractive. For us in local journalism in particular, it's currently not a question of putting employees out on the street, but of covering areas where we are short of people. We also face the challenge of being on site across the board in local journalism. If we use AI in the production process to shorten texts or prepare them for different channels, we can save more time. If we shorten documents such as local council minutes or provide reporters with a summary, they can recognize the exciting stories more quickly.

Where do you use AI to generate content?

Jakat: We are currently working on automating announcements from local institutions. For example, if a kindergarten party is taking place, the print edition needs a two-liner with the time, location and information on catering. There should be an input mask for this, in which kindergartens simply enter their information and an AI-generated text is created. We save resources here, but also improve quality because a potential transmission error has been eliminated.

Could it also be an advantage for media companies to strategically boycott AI?

Paganini: That doesn't make much sense. Throughout cultural history, there have always been strong emotions surrounding major media upheavals. Some are euphoric, others react in panic. Denial doesn't make sense because we live in precisely this world in which technological progress is taking place and which we should help to shape. It is our job to deal with reality, not to avoid it

Jakat: Of course, AI is also a challenge for us because we have to realize what our role actually is and how we do our job properly. We should talk about the value of our own research, i.e. original journalistic production, and what social and democratic added value we create as journalists. Curation has never really been what a media company is all about. If we are transparent and promote our work, there is enormous potential. Now is a good time to do so.