Search engines help us to find information on the Internet. Digital assistants answer our questions in a few sentences. We have to remember less because the knowledge is available at all times. But the systems have their pitfalls and are not always error-free. Do you check the answers or do you assume that they will be correct? Do you also look at the links that only appear on the second or third page? What do generative AI and digital assistants mean for how we deal with news, news sites and the search for knowledge? Deutsche Telekom is currently investigating this in a large survey with the Allensbach Institute. In the photo show, we show (already now) a few examples of why the topic is so relevant.
Endless link lists even for the simple search for potting soil
Actually, we only need one answer and not numerous. But search engines provide pages and pages of links. Let's be honest: Do you look at the links on the second or even third page? You are not alone: Studies show that only 40.3 clicks result from 100 search queries on Google. So almost 60 percent are satisfied with the answer at the top and do not check the source.
2024 Zero-Click Search Study
References need to be clicked
Tools like Perplexity advertise that we no longer have to search, but find. Instead of long lists of links, they provide us with answers. Well written, plausible-sounding and concise. With references. Hand on heart: How often do you click on the sources and check whether the answer is really correct?
Who determines what is?
After the assassination of Donald Trump on July 13, Meta AI initially had nothing to say about it. Meta explained in a blog that it had initially deliberately not let the chatbot talk about the attack. Because around headline-grabbing events, there are often contradictory statements and unfounded speculation up to conspiracy theories. On the one hand, understandable. But who determines what bots give answers to and what not?
Media and digital assistants
ZEIT is experimenting with an AI-based application to answer readers' questions about current events from its archive.
The Frankfurter Allgemeine Zeitung (FAZ) offers its subscribers an AI-generated summary of selected news articles in its news app.
Axel Springer is cooperating with OpenAI. ChatGPT is allowed to access selected content from Bild, Welt, Politico, and Business Insider to train ChatGPT to provide information about current events. For this purpose, ChatGPT links to the Springer articles in its answers.
Der Spiegel has entered into a three-year cooperation with the AI response bot Perplexity to make Spiegel articles more visible on this platform. In return, Spiegel receives a share of the advertising revenue generated with its content.
Opinions on AI in journalism
The use of AI is accepted if it helps media professionals to work more effectively (translations, transcription from speech to writing, etc.). According to a survey, on the other hand, only 14 percent of respondents in Germany are in favor of news that is mainly produced by AI.
However, the field of application makes a difference: the summary of a sporting event is more likely to be accepted than an article on political topics.
AI-generated and realistic-looking photos and videos are viewed particularly critically. This is because people want to "rely on images and videos as 'mental shortcuts' so they can see what they can trust online." True to the motto "Seeing is believing" or in German "Seeing is believing".
Regardless of the application, users believe that full automation should be taboo and that a human must always be involved a human must always be involved.
