Abstract Although Artificial Intelligence can offer significant business benefits, many consumers have negative perceptions of AI, leading to negative reactions when companies act ethically and disclose its use. Based on the pervasive example of content creation (e.g., via tools like ChatGPT), this research examines the potential for human-AI collaboration to preserve consumers' message credibility judgments and attitudes towards the company. The study compares two distinct forms of human-AI collaboration, namely AI-supported human authorship and human-controlled AI authorship, with traditional human authorship or full automation. Building on the compensatory control theory and the algorithm aversion concept, the study evaluates whether disclosing a high human input share (without explicit control) or human control over AI (with lower human input share) can mitigate negative consumer reactions. Moreover, this paper investigates the moderating role of consumers’ perceived morality of companies’ AI use. Results from two experiments in different contexts reveal that human-AI collaboration can alleviate negative consumer responses, but only when the collaboration indicates human control over AI. Furthermore, the effects of content authorship depend on consumers' moral acceptance of a company's AI use. AI authorship forms without human control lead to more negative consumer responses in case of low perceived morality (and no effects in case of high morality), whereas messages from AI with human control were not perceived differently to human authorship, irrespective of the morality level. These findings provide guidance for managers on how to effectively integrate human-AI collaboration into consumer-facing applications and advises to take consumers' ethical concerns into account.