Generative artificial intelligence (AI) tools have increasingly been reported to be deployed to generate, summarize, and personalize news content, increasingly mediating the production and circulation of information. They not only automate journalistic functions but also reconfigure the conditions under which knowledge is created, legitimized, and consumed. This epistemic shift raises concerns about how trust and authority are conferred in digital environments and how economic logic facilitates this transformation. This study aims to answer three central research questions: (1) How do generative AI systems construct themselves as epistemically authoritative sources of information? (2) How do these systems emerge as credible sources? (3) How do these systems frame news-related content in ways that commodify this credibility? Through a thematic discourse analysis of a sample of n = 30 items this study critically examines how AI has been constructing itself as an epistemically authoritative source, and how AI response framing aligns—or fails to align—with sustainable development goals. Findings suggest that Gemini-generated content particularly simulates epistemic authority through its visual presentation, downplaying epistemic transparency and ethical complexity. Credibility is engineered rather than organically earned, and commodification happens within Google’s vertical model. These findings have implications for understanding the transformation of trust in digital knowledge environments, raising critical concerns about AI’s role in reinforcing post-truth conditions, and the necessity of aligning AI practices with global sustainable and equitable goals.