How to edit a liquid: a survival guide for the AI age

As editors, our traditional model of responsibility is all about predictability and control. Read, ask for changes, repeat. Our professional practice is all about confidence in that final edit. Once we sign off, it’s sacred. Not a word may change, no extra tape can be cut in, no images resized. What we approve is what goes live.
No more.
As publishers experiment with generative AI to create public-facing AI experiences – summaries, semantic search, chatbots, voice – there’s now the possibility at least that the public no longer sees exactly what we signed off. Because generative AI, as content, is "liquid": personalised and probabilistic. We can edit the prompts. We cannot be certain of outcomes, so - how do we do our jobs? How do we approach the balance between usefulness for the individual and the need for editorial oversight to maintain trust?
This panel will explore how newsrooms must redefine the editor's role for a dynamic, AI-driven future. Speakers come from some of Europe and the U.S.’s leading journalism brands, and will draw on practical experience navigating the tension between instant, hyper-personalized AI services and the immutable, essential requirement for factual accuracy. The discussion will delve into the challenges presented by generative AI’s inability to guarantee deterministic outputs and the possibility of shifting focus from moderating individual pieces to managing models and overall user experiences.
Session takeaways:
> How generative AI challenges traditional legal and ethical models of editorial responsibility.
> Strategies for defining acceptable error tolerance in AI-driven, personalized news products.
> Practical steps newsroom leaders can take to workshop new principles.
> The shift from publishing a "pre-moderated product" to managing a "dynamic service."
> How to ensure consistent editorial oversight.
This session is essential for anyone who is proactively working to ensure that journalistic integrity and accountability remain central in a future defined by algorithmic content creation and distribution.
Moderated by Olle Zachrison.
Organised in association with The New York Times.