The Guardian has published three broad principles governing its use of generative AI (GenAI) tools, following a three-month review by a cross-departmental working group. The principles aim to balance the potential benefits of the technology with the need to maintain journalistic integrity and reliability.
Under the first principle, the Guardian will only use GenAI editorially where it contributes to the creation or distribution of original journalism, with human oversight and control. Any significant AI-generated elements in a piece of work will require clear evidence of benefit, senior editor approval, and transparency with readers.
The second principle focuses on improving the quality of work, such as helping journalists analyse large data sets, assisting with corrections or suggestions, or reducing bureaucratic processes. The Guardian emphasises that its core mission remains serious reporting that uncovers facts and holds power to account.
Finally, the Guardian will consider the ethical implications of AI models, including permissioning, transparency, and fair reward for creators. The organisation notes that its investment in journalism generates revenue through licensing, and using GenAI does not waive rights in its underlying content.
The Guardian, founded 202 years ago, has a history of adapting to technological change. It believes trusted media organisations that prioritise original reporting will remain essential, even as the full impact of AI on society is yet to be understood.
