There is diminishing trust over published research articles due to inaccurate, mis leading, or fraudulent content, which can be enhanced by use of generative AI tools in the research or writing process without appropriate guardrails. Although many initiatives aim to increase trust, authorship integrity cannot be underestimated. Author ship comes with many responsibilities, and yet repeated authorship misconduct occurs either intentionally (e.g. buying authorship) or due to suboptimal training in the responsibilities a person must undertake to be called an author. Addressing the gaps in understanding ethical authorship practices is critical to fostering a healthy research integrity culture, and stakeholders across the research ecosystem play a role in ensuring clear understanding of authorship roles. This article explores from a publisher perspective the different types and possible reasons for authorship misconduct, and how improving authorship practices requires optimal training to be co-developed and deployed across stakeholders.