Increasingly, people are using AI assistants like ChatGPT, Gemini, Claude, or Grok to stay informed about current news. The literature shows that these systems select their sources selectively and with little transparency, which can have far-reaching implications for the topic, tone, political bias, or susceptibility to misinformation in news summaries. 

To protect consumers from biased information, comprehensive regulations are being discussed and partially implemented at both the German and European levels. For instance, the second part of the Digital Media Treaty (Digitale-Medien-Staatsvertrag) is set to introduce so-called "AI information systems" as a new category under media law for the first time. This would include transparency requirements and make providers accountable for the generated content. At the European level, the EU Commission classified ChatGPT as a very large online search engine under the Digital Services Act (DSA) in August 2026, meaning that from January 2027, ChatGPT will fall under the DSA's regulations. Additionally, the EU KIDS Act, proposed in September 2026, aims to prevent AI companions and chatbots from being designed in ways that cause minors to become emotionally dependent on them. 

To analyze and evaluate these regulatory measures, a solid data basis is necessary. Therefore, we are creating a time-series dataset from automated AI queries that documents news summaries and the sources used by various AI assistants three times a week. These queries are sent from different regions—USA, Germany, Switzerland. This continuity allows us to measure the effects of different regulatory measures, for example, regarding source selection, tone, political bias, and susceptibility to misinformation, in a causal manner. 

Furthermore, the infrastructure set up for automated questioning of AI assistants can be extended to cover additional, current topics, providing up-to-date data on policy-relevant questions in a short amount of time. This enables the derivation of well-founded recommendations for regulation, the promotion of media literacy, and the individual use of AI.