Summary
This July 15, 2024 peer-reviewed study analyzed 37 AI guidelines used by media organizations across 17 countries and found a recognizable global pattern: transparency, disclosure, human oversight, and bias control recur constantly, but adoption remains geographically uneven and many guidelines still leave operational details underspecified.
Why It Matters
For journalists and newsroom leaders, this is a durable direct governance reference because it compares how real media organizations were trying to operationalize AI rather than offering another abstract ethics manifesto.
- it shows that disclosure, human verification, and bias control became baseline expectations across many newsroom AI policies
- it identifies recurring concerns about hallucinated content, workflow disruption, misinformation, and the skills gap needed to supervise AI systems
- it documents a geographic concentration of guideline production in Europe and North America, which matters when people talk as if the industry's norms were already globally settled
- it helps separate broad principle statements from more operational guidance about how journalists and AI systems should interact in practice
What the Source Says
The paper says the corpus included 37 AI guidelines from media organizations worldwide. Its close reading found repeated concern about false or misleading AI-generated content, workflow disruption, disinformation, transparency, and the continued need for human verification. It also found that early guideline adoption was concentrated in North America and Europe, with later and narrower uptake elsewhere. The authors note that some organizations explicitly emphasize disclosure and four-eyes review, while others speak more vaguely about trust and safety without spelling out implementation detail.