Generative AI is simultaneously labelled a threat and a boon for data
journalists. Cultivating your own perspective requires strategic,
hands-on engagement in focused ways, not policy papers or uncritical
adoption. We must encourage ourselves, and those we teach, to center
learning experiments on our well-known journalistic tenets of
verification and transparency. Classroom and professional educators
are improvising with little shared understanding of what “AI
literacy” might mean for data journalists. Efforts to grow our
ability to use generative AI tools are outpacing our ability to know
when or when not to use them. This gap has led to numerous gaffes and
retractions, from media publications at different scales across the
globe. Do you trust AI to perform data classification tasks at scale?
Can mundane graphics-desk style charts be created automatically? Does
code generation enhance our ability to create dynamic interactive
visualizations? I argue that “old school” principles such as
verification and transparency are key to how we train ourselves and
the next generation to answer questions like these. Through examples
from various data-related tasks I’ll introduce you to some real
examples from US newsrooms and studies on AI adoption that link to
deeper conversations you’re likely to have in sessions throughout
the event. Together in spaces like this we can push forward our
ability to prepare ourselves for the use, misuse, and refuse of
generative AI in data journalism.