An increasing amount of news is generated automatically by artificial intelligence (AI). While the technology has advantages for content production, e.g., regarding efficiency in aggregating information, it is also viewed critically due to little transparency in obtaining results and possible biases. As news media are dependent on trust and credibility, introducing AI to facilitate mass communication with consumers seems to be a risky endeavor. We expand research on consumer perception of AI-based news by comparing machinewritten and human-written texts to fake news and by examining the role of trust that consumers exhibit when evaluating news. Through an experiment with 263 participants, we find that consumers judge AI-based texts similar to true journalistic content when it comes to credibility, but similar to fake news regarding readability. Furthermore, our results indicate that consumers with low trust in media are less averse to AIbased texts than consumers with high trust in media.