Power, voice, and accountability: how artificial intelligence reshapes inclusive science communication in evidence-based healthcare—a narrative literature review

Inclusive science communication is increasingly recognised as essential to evidence-based healthcare, yet the concept remains contested, and practical guidance for operationalising it is limited. The rapid integration of artificial intelligence into health communication transforms these unresolved tensions from theoretical concerns into practical imperatives, as AI systems increasingly make consequential decisions about whose evidence is prioritised, how it is framed, and who can access it. This narrative literature review synthesises 79 sources across science communication, health communication, AI ethics, and evidence-based healthcare to analyse competing conceptualisations of inclusive science communication, examine how AI reshapes the challenges of voice, access, accountability, and power, and evaluate frameworks that can inform equity-centred practitioner guidance. We identify three defining tensions within the field and derive six assessable dimensions of inclusive practice. We then demonstrate that AI reconfigures the mechanisms by which exclusion operates, embedding communicative decisions within opaque systems that outpace existing oversight. Evaluating existing frameworks against these dimensions, we find a persistent structural gap: no current framework or tool integrates quality and equity assessment or addresses AI-specific challenges for inclusive practice. We conclude that integrated guidance treating quality and inclusion as interdependent is needed and must be co-developed with the communities whose health outcomes are at stake.